No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Signal Processing and Machine Learning Techniques for Terahertz Sensing: An overview
Abstract
Following the recent progress in terahertz (THz) signal generation and radiation methods, joint THz communications and sensing (CAS) applications are being proposed for future wireless systems. Toward this end, THz spectroscopy is expected to be carried over user equipment devices to identify material and gaseous components of interest. THz-specific signal processing techniques should complement this resurgent interest in THz sensing for efficient utilization of the THz band. In this article, we present an overview of these techniques, with an emphasis on signal preprocessing [standard normal variate (SNV) normalization, minimum–maximum normalization, and Savitzky–Golay (SG) filtering], feature extraction [principal component analysis (PCA), partial least squares (PLS),
t
-distributed stochastic neighbor embedding (
t
-SNE), and nonnegative matrix factorization (NMF)], and classification techniques [support vector machines (SVMs), the
k
-nearest neighbor (
k
NN), discriminant analysis (DA), and naive Bayes (NB)]. We also address the effectiveness of deep learning techniques by exploring their promising sensing and localization capabilities at the THz band. Finally, we investigate the performance and complexity tradeoffs of the studied methods in the context of joint CAS (JCAS). We thereby motivate corresponding use cases and present a handful of contextual future research directions.
... Also, many chemical and biological materials exhibit unique spectral fingerprinting in the THz spectrum for example the vibrational modes of THz allow molecular structure and vibrational dynamics studies that arise from intermolecular and intramolecular interactions. Further, the THz radiation is strongly absorbed by water molecules which enable water dynamics observations and inspection of metallic components enabled through the capability of THz radiation to be reflected by metals and penetrate dielectric, amorphous and non-conducting materials [2]. ...
... The reconstruction and denoising tasks which are intended for image and spectrum preprocessing to remove unwanted or irrelevant information and reduce number of variables from spectral and image parameters thus increasing the data analysis efficiency [4]. The machine learning methods are also used for multivariate quantitative and qualitative analysis of data for highly precise classification and recognition of samples [2]. ...
... An in-depth review of machine learning and deep learning techniques for signal processing and classification for efficient THz communication and sensing was presented in [2]. The article promoted the significance of THz frequency domain spectroscopy (FDS) and THz time domain spectroscopy (THz TDS) in future reconfigurable. ...
Following the recent progress in the development of Terahertz (THz) generation and detection, THz technology is being widely used to characterize test sample properties in various applications including nondestructive testing, security inspection and medical applications. In this paper, we have presented a broad review of the recent usage of artificial intelligence (AI) particularly, deep learning techniques in various THz sensing, imaging, and spectroscopic applications with emphasis on their implementation for medical imaging of cancerous cells. Initially, the fundamentals principles and techniques for THz generation and detection, imaging and spectroscopy are introduced. Subsequently, a brief overview of AI – machine learning and deep learning techniques is summarized, and their performance is compared. Further, the usage of deep learning algorithms in various THz applications is reported, with focus on metamaterials design and classification, detection, reconstruction, segmentation, parameter extraction and denoising tasks. Moreover, we also report the metrics used to evaluate the performance of deep learning models and finally, the existing research challenges in the application of deep learning in THz cancer imaging applications are identified and possible solutions are suggested through emerging trends. With the continuous increase of acquired THz data – sensing, spectral and imaging, artificial intelligence has emerged as a dominant paradigm for embedded data extraction, understanding, perception, decision making and analysis. Towards this end, the integration of state-of-the-art machine learning techniques such as deep learning with THz applications enable detailed computational and theoretical analysis for better validation and verification than modelling techniques that precede the era of machine learning. The study will facilitate the large-scale clinical applications of deep learning enabled THz imaging systems for the development of smart and connected next generation healthcare systems as well as provide a roadmap for future research direction.
... The ratio between the transmitted and received signals illustrates the material properties. Specifically, two varieties of THz-TDS techniques are available to measure the aforementioned properties: transmission and reflection spectroscopy [4], [7]. Here, transmission spectroscopy refers to measuring the amount of light absorption, whereas reflection spectroscopy measures light reflection or scattering. ...
... Several studies explore material identification using conventional feature extraction techniques and deep learning [7], [13]. A material identification using approximate entropy and deep neural network (DNN) is presented in [13], where 14 different materials are investigated for identification. ...
... Also, the proposed approach shows 80.4% accuracy in detecting the materials without noise. A recent study on THz sensing based on signal processing and ML is presented in [7], where pre-processing, feature extraction, and performance analysis of THz spectroscopy are discussed. Correspondingly, authors in [7] compared different pre-processing techniques and existing ML algorithms for 5 materials. ...
p>Terahertz (THz) spectrum is identified as a potential enabler for advanced sensing and positioning, where THz-Time domain spectroscopy (THz-TDS) is specified for investigating the unique material properties. The transmission THz-TDS measures the light absorption of materials. This paper proposes a novel low-complex deep neural network (DNN)-based multi-class classification architecture to sense a wide variety of materials from the transmission spectroscopy. Based on the spectroscopic measurements made across a chosen THz region of interest, DNN extracts and learns the distinctive crystal structure of materials as features. With sufficient quantities of noisy spectroscopic data and labels, we train and validate the model. In low SNR regions, the proposed DNN classification architecture achieves about 92%success rate, which is greater than those of the state-of-the-art methods.</p
... The ratio between the transmitted and received signals illustrates the material properties. Specifically, two varieties of THz-TDS techniques are available to measure the aforementioned properties: transmission and reflection spectroscopy [4], [7]. Here, transmission spectroscopy refers to measuring the amount of light absorption, whereas reflection spectroscopy measures light reflection or scattering. ...
... Several studies explore material identification using conventional feature extraction techniques and deep learning [7], [13]. A material identification using approximate entropy and deep neural network (DNN) is presented in [13], where 14 different materials are investigated for identification. ...
... Also, the proposed approach shows 80.4% accuracy in detecting the materials without noise. A recent study on THz sensing based on signal processing and ML is presented in [7], where pre-processing, feature extraction, and performance analysis of THz spectroscopy are discussed. Correspondingly, authors in [7] compared different pre-processing techniques and existing ML algorithms for 5 materials. ...
Terahertz (THz) spectrum is identified as a potential enabler for advanced sensing and positioning, where THz-Time domain spectroscopy (THz-TDS) is specified for investigating the unique material properties. The transmission THz-TDS measures the light absorption of materials. This paper proposes a novel low-complex deep neural network (DNN)-based multi-class classification architecture to sense a wide variety of materials from the transmission spectroscopy. Based on the spectroscopic measurements made across a chosen THz region of interest, DNN extracts and learns the distinctive crystal structure of materials as features. With sufficient quantities of noisy spectroscopic data and labels, we train and validate the model. In low SNR regions, the proposed DNN classification architecture achieves about 92%success rate, which is greater than those of the state-of-the-art methods.
... In the context of millimeter wave operations, the available bandwidth options typically span from 30 to 300 GHz [1]. However, the advent of Sixth Generation (6G) networks necessitates a greater allocation of radio resources, which entails a shift towards utilizing terahertz frequency ranges [2]. One of the salient attributes of millimeter wave technology is its ability to support high data rates for 6G networks, even at low frequencies [3]. ...
... In order to build terahertz communication, it is important to mitigate a greater number of undesirable signals from 6G networks, hence posing a significant challenge to current generation networks. Furthermore, the integration of terahertz communication systems with millimeter waves can be achieved, wherein the processing of each state is effectively managed through the utilization of machine learning algorithms [2]. The utilization of machine learning techniques enables the comprehensive observation of entire information pertaining to all states through the usual variation process. ...
The proliferation of integrated sensing techniques in Sixth Generation (6G) networks is an increasingly significant aspect in facilitating efficient end-to-end communication for all users. The suggested methodology employs a digital signal processed with terahertz bandwidth to assess the impact of 6G networks. The primary focus lies in the design of 6G networks, emphasizing key parameters such interference, loss, signal strength, signal-to-noise ratio, and dual band channels. The aforementioned factors are combined with two machine learning algorithms in order to determine the extent of spectrum sharing among all available resources. Thus suggested approach for detecting signals in the terahertz communication spectrum is evaluated using 10 devices across four situations, which involve interference, signal loss, strength, and time margins for integrated sensing. Also the assumptions are based on signal processing devices operating within millimeter waves ranging from 5 to 10 terahertz. Interference and losses in the specified spectrum are seen to be less than 1%, but the time margin for integrated sensing with 99% maximized signal intensity remains at 85%.
... The application of machine learning techniques in THz imaging and sensing have been extensively reviewed in [181], [182], [183], [184], [185], [186], [187] and [188]. However, most of the machine learning models that have been explored in THz imaging and sensing for biomedical applications are based on shallow networks due to the unavailability of sufficient training datasets. ...
... In the future of THz technology development, more advanced and robust machine learning algorithms will be required that are capable of finding accurate and reasonable solutions in circumstances of out of distribution data and far from learned distribution data. Such techniques that will benefit future THz sensing and communication applications include deep learning, multitask learning, meta learning, federated learning, active learning, specialized learning [181], [183]. ...
There is a keen interest in the exploration of new generation emitters and detectors due to advancements in innovation of new materials and device processing technologies which have opened up new frontiers in the Terahertz (THz) spectrum. Therefore, it is necessary to review the developments in THz technology for healthcare applications, their impact, implications and prospects for ongoing research and development. This paper provides a broad overview of the current status and prospects of application of THz imaging and sensing for the healthcare domain. We present current knowledge, identify existing challenges for wide scale clinical adoption of THz systems and prospective opinions to facilitate research and development towards optimized and miniaturized THz systems and biosensors that provide real operational convenience through emerging trends. Firstly, we provide an overview of the THz imaging and sensing techniques that exploit properties of THz generation and detection with emphasis on terahertz time domain spectroscopy (THz-TDS) and THz Metamaterials. The mechanisms of tissue image contrast and the application of THz imaging and sensing for biomedical applications in particular, the cancer detection application is reported. Secondly, an outlook toward the advancements in THz technology in the interface of healthcare 4.0 and its enabling technologies is explored for next generation smart and connected healthcare systems. Third, we identify the merits and existing challenges in THz cancer imaging and sensing and suggest prospective opinions to pave way to ongoing and future research. Further, we discuss the recent advances in THz imaging development and the contribution of near-field techniques based on plasmonic, and resonance based metasurfaces, waveguides etc. for breaking the diffraction limit towards development of THz systems that are convenient for point of care. We bring researchers a roadmap for future research scope.
... © [2022] IEEE. Reprinted, with permission, from[136]. ...
The development of 6 G networks has promoted related research based on terahertz communication. As submillimeter radiation, signal transportation via terahertz waves has several superior properties, including non-ionizing and easy penetration of non-metallic materials. This paper provides an overview of different terahertz detectors based on various mechanisms. Additionally, the detailed fabrication process, structural design, and the improvement strategies are summarized. Following that, it is essential and necessary to prevent the practical signal from noise, and methods such as wavelet transform, UM-MIMO and decoding have been introduced. This paper highlights the detection process of the terahertz wave system and signal processing after the collection of signal data.
... Imaging in the terahertz band (0.1 to 10 THz, equivalent to 3 mm to 30 μm in wavelength) has seen great growth in interest in recent years due to its nonionizing nature as well as the high degree of transparency in this spectral regime in many nonpolar dielectrics (1)(2)(3). Terahertz time-domain spectroscopy (THz-TDS) has become a leading nondestructive spectroscopic testing and 3D imaging technique for determining the properties of a sample probed by short pulses of THz electromagnetic radiation (4,5). ...
Noninvasive inspection of layered structures has remained a long-standing challenge for time-resolved imaging techniques, where both resolution and contrast are compromised by prominent signal attenuation, interlayer reflections, and dispersion. Our method based on terahertz (THz) time-domain spectroscopy overcomes these limitations by offering fine resolution and a broadband spectrum to efficiently extract hidden structural and content information from layered structures. We exploit local symmetrical characteristics of reflected THz pulses to determine the location of each layer, and apply a statistical process in the spatiotemporal domain to enhance the image contrast. Its superior performance is evidenced by the extraction of alphabetic characters in 26-layer subwavelength papers as well as layer reconstruction and debonding inspection in the conservation of Terra-Cotta Warriors. Our method enables accurate structure reconstruction and high-contrast imaging of layered structures at ultralow signal-to-noise ratio, which holds great potential for internal inspection of cultural artifacts, electronic components, coatings, and composites with dozens of submillimeter layers.
... Complementary machine learning techniques that classify materials based on their spectral absorption coefficients in the THz range can play an important role. Helal et al. have reviewed and summarised different techniques that can potentially be used for THz signal processing [193]. They focused on the following classification techniques: ...
The increasing scarcity of natural resources, worsening global climate change, environmental degradation, and rising demand for food are forcing the biotechnology and plastics industries to seek and apply circular economy models that would lead to a sustainable transition in the production and use of bioplastics. Circular economy models can improve the economic productivity of bio-based plastics and have a positive impact on the environment by reducing conventional plastic waste and the consumption of petrochemical feedstocks for plastic production. In addition, some agricultural wastes that have the potential to be used as bioplastics can be reused. Terahertz (THz) systems are already used in the plastics and rubber industries for non-destructive testing, detection, imaging, and quality control. Several reports have highlighted the potential applications of THz spectroscopy and imaging in polymer analysis and plastics characterisation. This potential is even greater with chemometric methods and artificial intelligence algorithms. In this review, we focus on applications that support the transformation of the biotechnology sector to the circular economy, particularly via the transition from conventional plastics to bioplastics. In this review, we discuss the potential of THz systems for the characterisation and analysis of bioplastics and biopolymers. The results of previous studies on biopolymers in the THz frequency range are summarised. Furthermore, the potential of using artificial intelligence approaches such as machine learning as advanced analytical methods in THz spectroscopy and imaging, in addition to the conventionally used chemometric methods, is discussed. The results of this review highlight that THz technology can contribute to closed technological circles in important areas of biotechnology and the related plastics and rubber industries.
... Terahertz spectroscopy and imaging techniques are widely used in various fields for sample detection, and the acquired terahertz signals containing a large amount of sample information need to be analyzed in order to have a clearer understanding of the sample detection information. The combination of terahertz techniques with chemometric methods, machine learning, and search algorithms can effectively build analytical models of sample information [62]. In recent years, machine learning techniques have developed rapidly, and many new learning models and learning algorithms have emerged. ...
The frequency range of terahertz waves (THz waves) is between 0.1 and 10 THz and they have properties such as low energy, penetration, transients, and spectral fingerprints, which are especially sensitive to water. Terahertz, as a frontier technology, have great potential in interpreting the structure of water molecules and detecting biological water conditions, and the use of terahertz technology for water detection is currently frontier research, which is of great significance. Firstly, this paper introduces the theory of terahertz technology and summarizes the current terahertz systems used for water detection. Secondly, an overview of theoretical approaches, such as the relaxation model and effective medium theory related to water detection, the relationship between water molecular networks and terahertz spectra, and the research progress of the terahertz detection of water content and water distribution visualization, are elaborated. Finally, the challenge and outlook of applications related to the terahertz wave detection of water are discussed. The purpose of this paper is to explore the research domains on water and its related applications using terahertz technology, as well as provide a reference for innovative applications of terahertz technology in moisture detection.
... There is need for development of more robust and advanced machine learning algorithms with the capability to accurately find reasonable solutions in the cases where there is far from learned distribution data and out of distribution data. Some of the advanced machine learning techniques with great potential to transform the future of THz cancer imaging and sensing are for example deep learning, federated learning, active learning, meta-learning, multitask and specialized learning (Helal et al. 2022;Jiang et al. 2022). ...
There has been a rapid development of THz technology—sources, detectors and various THz imaging and sensing techniques. The THz technology demonstrates great potential as a modality for early, label free, non-ionizing and non-invasive detection of cancer. Some progressive technological development milestones have been achieved in this regard, however, to become clinically competitive and to provide the sought after real operational convenience, there is need for further research and development to overcome the existing challenges. This paper provides recent trends and perspectives through identification of existing challenges for the development of THz imaging and sensing systems that can evolve into actual medical modalities. We provide an overview of various aspects of THz technology, including techniques for imaging and sensing, mechanisms for THz image contrast and models for tissue dielectric responses to THz waves. The THz imaging application for detection of various cancers is briefed. The advantages of THz cancer imaging and sensing as well as the existing challenges are identified, with recommendations provided in contribution to future research. Further, some recent THz imaging and sensing developments such as the near-field methods to break the diffraction limit including waveguides, resonance and plasmonic metasurfaces are discussed. We emphasize the contribution of analytical algorithms that are based on machine learning, in particular, deep learning for the development of THz technology.
Graphical abstract
... Among these methods, NMF has shown good performances due to its ability of producing a parts-based representation. Therefore, NMF-based methods have widely used in numerous actual tasks such as document analysis, bioinformatics, information retrieval and so on [6][7][8][9][10][11]. ...
Multiple graph and semi-supervision techniques have been successfully introduced into the nonnegative matrix factorization (NMF) model for taking full advantage of the manifold structure and priori information of data to capture excellent low-dimensional data representation. However, the existing methods do not consider the sparse constraint, which can enhance the local learning ability and improve the performance in practical applications. To overcome this limitation, a novel NMF-based data representation method, namely, the multiple graph adaptive regularized semi-supervised nonnegative matrix factorization with sparse constraint (MSNMFSC) is developed in this paper for obtaining the sparse and discriminative data representation and increasing the quality of decomposition of NMF. Particularly, based on the standard NMF, the proposed MSNMFSC method combines the multiple graph adaptive regularization, the limited supervised information and the sparse constraint together to learn the more discriminative parts-based data representation. Moreover, the convergence analysis of MSNMFSC is studied. Experiments are conducted on several practical image datasets in clustering tasks, and the clustering results have shown that MSNMFSC achieves better performance than several most related NMF-based methods.
... The work in [107] discussed the development of ML techniques and proposed an intelligent system with the ability to perceive the environment, evaluate the resources/spectrum and learn how to adapt system configurations to make the most use of the given resources. The study in [108] reviewed signal processing and ML Techniques for THz JSAC systems. Overall, ML is a strong tool that can optimize efficient wireless resource utilization in JSAC since resource allocation optimization issues are typically too difficult to model owing to dynamic wireless settings. ...
Since the 1960s, joint sensing and communication (JSAC) has been proposed as an attractive technique with advantages of enhanced spectral and hardware efficiency along with low latency. However, in those old days, complex transceiver designs hindered the massive adoption of JSAC in many applications. Nevertheless, thanks to advancing wireless technologies in recent years, JSAC has recently attracted substantial attention for a wide range of civil and military applications. In particular, JSAC enables simultaneous sensing and communication functionalities with the full cooperation of both operations in shared resources such as hardware as well as radio resources (i.e., frequency, time, space and so on). Note that sensing functionality is associated with other sub-functionalities of radio-detection-and-ranging (Radar), computation, and localization which are sporadically used terms in the literature. Thus, to generalize the concept and avoid any confusion, we use the term “sensing” in the acronym JSAC as a consistent and general term associated with the other aforementioned sub-functionalities. This paper elaborates on defining sensing and communication operations and then, provides an overview with preliminaries, key latest findings, and state-of-the-art of JSAC. It also explores both existing and emerging Internet of Things (IoT) applications of JSAC. Next, it provides a new classification of JSAC technologies by taking into account not only the existing JSAC technologies but also a diverse range of technologies that allow JSAC to be used for various types of IoT applications. Eventually, this study projects future research directions and challenges of enabling JSAC in IoT.
... The work in [108] discussed the development of ML techniques and proposed an intelligent system with the ability to perceive the environment, evaluate the resources/spectrum and learn how to adapt system configurations to make the most use of the given resources. The study in [109] reviewed signal processing and ML Techniques for THz JSAC systems. Overall, ML is a strong tool that can optimize efficient wireless resource utilization in JSAC since resource allocation optimization issues are typically too difficult to model owing to dynamic wireless settings. ...
This article is under review and upon acceptance: Personal use of this material is permitted. However, permission to use this material for any other purposes must be obtained from the IEEE by sending a request to pubs-permissions@ieee.org. (Copyright (c) 2015 IEEE.)
... The work in [108] discussed the development of ML techniques and proposed an intelligent system with the ability to perceive the environment, evaluate the resources/spectrum and learn how to adapt system configurations to make the most use of the given resources. The study in [109] reviewed signal processing and ML Techniques for THz JSAC systems. Overall, ML is a strong tool that can optimize efficient wireless resource utilization in JSAC since resource allocation optimization issues are typically too difficult to model owing to dynamic wireless settings. ...
This article is under review and upon acceptance: Personal use of this material is permitted. However, permission to use this material for any other purposes must be obtained from the IEEE by sending a request to pubs-permissions@ieee.org. (Copyright (c) 2015 IEEE.)
... Detecting concealed objects underneath clothing is an essential step in public security checks, while the manual check is criticized for inefficiency, invasion of privacy, and high rate of missed detection. Terahertz band, between microwave and infrared, with the frequency range from 0.1 to 10 Terahertz [8], provides a non-contact way to discover objects concealed underneath clothing with no harm to human health. According to the presence or absence of Terahertz source irradiation, there are two categories of Terahertz imaging systems -passive [9], [20], [16] and active [3], [19], [10]. ...
Concealed object detection in Terahertz imaging is an urgent need for public security and counter-terrorism. In this paper, we provide a public dataset for evaluating multi-object detection algorithms in active Terahertz imaging resolution 5 mm by 5 mm. To the best of our knowledge, this is the first public Terahertz imaging dataset prepared to evaluate object detection algorithms. Object detection on this dataset is much more difficult than on those standard public object detection datasets due to its inferior imaging quality. Facing the problem of imbalanced samples in object detection and hard training samples, we evaluate four popular detectors: YOLOv3, YOLOv4, FRCN-OHEM, and RetinaNet on this dataset. Experimental results indicate that the RetinaNet achieves the highest mAP. In addition, we demonstrate that hiding objects in different parts of the human body affect detection accuracy. The dataset is available at https://github.com/LingLIx/THz_Dataset.
... Beyond signal acquisition, several signal processing and machine learning techniques can be used to pre-process the received signals, extract characteristic features, and classify target materials into appropriate classes [453]. Furthermore, the accuracy of sensing and imaging is greatly enhanced in the THz band due to the vastly wider available channel bandwidths and the high directionality that accompanies massive MIMO beamforming. ...
Terahertz (THz)-band communications are a key enabler for future-generation wireless communication systems that promise to integrate a wide range of data-demanding applications. Recent advancements in photonic, electronic, and plasmonic technologies are closing the gap in THz transceiver design. Consequently, prospect THz signal generation, modulation, and radiation methods are converging, and the corresponding channel model, noise, and hardware-impairment notions are emerging. Such progress paves the way for well-grounded research into THz-specific signal processing techniques for wireless communications. This tutorial overviews these techniques with an emphasis on ultra-massive multiple-input multiple-output (UM-MIMO) systems and reconfigurable intelligent surfaces, which are vital to overcoming the distance problem at very high frequencies. We focus on the classical problems of waveform design and modulation, beamforming and precoding, index modulation, channel estimation, channel coding, and data detection. We also motivate signal processing techniques for THz sensing and localization.
Millimeter-wave (mmWave) technology has improved to overcome propagation issues and enable high data speeds in 5G and 6G wireless networks. The capabilities of 5G networks are limited in their ability to support future applications such as extended reality and 3D gaming. Terahertz (THz) transmission, a key component of 6G networks that enable fast data transmission over short distances, has consequently attracted increasing attention. Addressing challenges, including signal interference, antenna design, regulatory compliance, and security concerns, is essential for realizing the full capabilities of wireless communication in the mmWave and THz frequency ranges.The Terahertz Channel Analysis Framework for Next-Generation Wireless Networks (TCAF-NGWN) examines complex design components, channel models, and signal propagation in the mmWave and THz bands. The report evaluates these bands' capabilities to fulfill 6G networks' evolving communication needs and examines commercial application standardization efforts. 6G terahertz communication channel research mainly occurs in academic labs and focuses on antenna arrays, channel properties, and transmitter/receiver combinations. Signal transmission is analyzed for route loss, multipath fading, signal attenuation, reflections, obstacles, and diffraction. Channel capacity, data rate, reflection coefficients, and signal-to-noise ratio are evaluated. The study analyses the theoretical models and improves communication systems using literature and datasets like the Terahertz Wireless Channel Dataset. Comparing the suggested model to THz-IRS, UM-MIMO, and CDOT shows its advantages. The research advances terahertz wireless communication technologies by providing insights into channel propagation and system performance in practical circumstances through precise experimental setups and measuring techniques.
The application of deep learning algorithms is rapidly expanding, Deep learning algorithms show great potential to revolutionize healthcare towards the development of smart clinical decision support systems. The development of computer-aided diagnostics systems is crucial to support clinicians with second diagnostic opinions, overcoming the subjectivity of manual diagnosis and handcrafted features. While significant progress has been made in multimodal image data fusion, the application of deep learning for processing of healthcare image and acoustic information has not yet been fully explored. It is through multisource information fusion and computational modeling that outcomes of interest such as treatment targets and drug development ultimately facilitate improved patient-level decision making in care facilities and homes. Such a phenomenon has attracted interest in healthcare multisource data fusion studies. The application of machine learning and deep learning algorithms provides insight into various outcomes of healthcare such as drug discovery, clinical trials, phenotyping, and surgical techniques. This is crucial in providing support to practitioners and health centers to provide the most precise and efficient evidence-based medicine possible. Moreover, unimodal deep learning models are often less robust and face misclassification challenges. In this chapter we focus on multisource medical information processing for disease classification and prediction tasks. Specifically, we propose a framework and algorithms for multimodal deep learning models for classifying acoustic and image type multimodal datasets for lung cancer. First, the data fusion techniques are reported. Second, a deep learning-based multimodal data fusion framework is proposed for multisource image and acoustic data processing. Further, the existing challenges in multisource information processing systems based on deep learning are identified and a prospective opinion is given. This will pave the way for future research.
The prognosis of organic light-emitting diodes (OLEDs) not only requires early detection of a bearing defect, but also the capability to predict their life data under all operational scenarios. The use of sophisticated machine learning (ML) algorithms is undoubtedly becoming an increasingly exciting research direction, as these algorithms can yield high predictive models with minimal domain expertise. The central question of this perspective is: how well can ML models advance our ability to forecast the lifetime of OLEDs compared to the physics-based models? In this paper, data-driven methods, feed-forward neural networks (FFNN), support vector machines (SVMs), k-nearest neighbors (KNNs), partial least squares regression (PLSR), and decision trees (DTs), are used to predict the lifetime and reliability of OLEDs through analyzing the lumen degradation data collected from the accelerated lifetime test. The final predicted results indicate that both the data-driven and our physics-based OLED lifetime models fit well the experimental data. The main drawback of the former method is that their efficacy is highly contingent on the quantity and quality of the operational dataset. Among all these methods, much more reliability information (time to failure) and the highest prediction accuracy can be achieved by FFNN.
The terahertz (0.1-10 THz) band holds enormous potential for supporting unprecedented dat
<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a</sup>
<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ra</sup>
tes and millimeter-level accurate sensing thanks to its ultra-broad bandwidth. Terahertz integrated sensing and communication (ISAC) is viewed as a game-changing technology to realize connected intelligence in 6G and beyond systems. In this article, challenges from the THz channel and transceiver perspectives, as well as difficulties of ISAC, are elaborated. Motivated by these challenges, THz ISAC channels are studied in terms of channel types, measurement, and models. Moreover, four key signal processing techniques to unleash the full potential of THz ISAC are investigated, namely, waveform design, receiver processing, narrowbeam management, and localization. Quantitative studies demonstrate the benefits and performance of the state-of-the-art signal processing methods. Finally, open problems and potential solutions are discussed.
Terahertz (THz) technology has firmly established itself as an effective sensing and nondestructive testing technique for the detection of substances and physico-chemical evaluation of materials and structural systems since its first emergence almost three decades ago. To date, both the effectiveness and accuracy of this technology have been extensively demonstrated in a myriad of applications across the spectrum of research and development all the way to process analytical technology, quality control, nondestructive testing, and structural health monitoring. These applications are generally enabled by the production and availability of advanced, versatile, robust, highly accurate, and industrially rugged THz spectroscopy and imaging systems, the unique properties of THz waves compared with other electromagnetic waves, as well as the advancements in electronics, photonics, and THz metamaterial systems development. This article presents a comprehensive state-of-the-art and state-of-the-practice review of sensing and nondestructive testing applications of THz technology and analyzes the role of THz metamaterials in enhancing the resolution and sensitivity of THz systems. The study also provides a general overview of the fundamentals of THz spectroscopy and imaging systems and discusses the suitability of THz sensing and nondestructive testing in a variety of real-world application scenarios (
e.g
. composites’ defect detection and evaluation, paints and coatings thickness measurement and characterization, biomolecule detection,
etc
.). Aspects such as the noise caused by the presence of barriers, challenges with experimental implementations and operability of THz systems, long times required to acquire THz images, as well as limited customizability and portability of currently available THz systems are also discussed.
Reparos compósitos de engenharia representam uma importante forma de reabilitação da integridade estrutural de regiões com perda de espessura, também sendo capazes de promover estanqueidade em defeitos tipo furo passante, em decorrência do processo de corrosão. Esta técnica de reparo também confere proteção anticorrosiva externa e apresenta a importante vantagem de ser depositada a frio, o que impulsionou sua aplicação de forma crescente em sistemas offshore. Tais reparos são dimensionados para cada caso de uso e têm uma vida útil projetada, após a qual, devem ser repostos por outros. No entanto, operando sob as severas condições ambientais offshore, aliadas às condições de operação das tubulações, o material compósito que constitui o reparo pode sofrer alterações microestruturais que alterem suas propriedades mecânicas e, consequentemente, sua vida residual em serviço. Visando analisar o efeito da névoa salina sobre tais materiais, amostras de dois fabricantes representativos foram submetidas a diferentes tratamentos de envelhecimento, com variação de temperatura e tempo, e posteriormente avaliadas através de ensaios não destrutivos. Por seu caráter de não contato, facilidade de inspecionar não metálicos e capacidade de caracterizar materiais submetidos a pequenas alterações estruturais, as ondas terahertz foram adotadas como uma metodologia de estudo neste projeto. Como abordagem inicial, após o pré-tratamento e padronização dos dados, os mesmos foram reduzidos em sua dimensionalidade via PCA (Principal Component Analysis) e submetidos a uma classificação por outros métodos de Aprendizado de Máquina. Os resultados mostraram que a técnica de terahertz foi capaz de distinguir com acerto as diferentes amostras em função da temperatura e do tempo de envelhecimento, indicando que a caracterização não destrutiva do reparo compósito no campo é viável e pode responder quanto a permanência do reparo em operação ou não. Os principais benefícios desta abordagem seriam a garantia operacional dos sistemas de tubulação offshore de forma segura e a otimização de recursos com a troca de reparos.
Terahertz (THz) is a promising technology for future wireless communication networks, particularly for 6G and beyond. The ultra-wide THz band, ranging from 0.1 to 10 THz, can potentially address the limited capacity and scarcity of spectrum in current wireless systems such as 4G-LTE and 5G. Furthermore, it is expected to support advanced wireless applications requiring high data transmission and quality services, i.e., terabit-per-second backhaul systems, ultra-high-definition streaming, virtual/augmented reality, and high-bandwidth wireless communications. In recent years, artificial intelligence (AI) has been used mainly for resource management, spectrum allocation, modulation and bandwidth classification, interference mitigation, beamforming, and medium access control layer protocols to improve THz performance. This survey paper examines the use of AI in state-of-the-art THz communications, discussing the challenges, potentials, and shortcomings. Additionally, this survey discusses the available platforms, including commercial, testbeds, and publicly available simulators for THz communications. Finally, this survey provides future strategies for improving the existing THz simulators and using AI methods, including deep learning, federated learning, and reinforcement learning, to improve THz communications.
In recent years, spectral analysis methods have developed rapidly. A key feature is the use of chemometric methods to process spectral data for performing qualitative and quantitative analysis of complex mixtures. The coupling of chemometric methods to spectroscopic techniques led to distinct advantages in speed, cost, efficiency, automation, and portability compared to the traditional methods in agriculture, food, pharmaceutical, petroleum, chemical, environmental, and medical fields. This paper comments on the review papers published during the past three years (2020–2022) on the topic of the combination of spectral and chemometric methods. The development status, existing challenges, and the future direction of this field is discussed.
The combination of terahertz (THz) spectroscopic measurements and multivariate calibration techniques has become a well-established technique in many research fields. However, intentional or unintentional changes in environmental conditions, THz instruments and/or of the substance itself make the established calibration model becoming insufficient and inadequate for the further application. In this article, we introduce, discuss, and evaluate a new multivariate calibration method, the CWT-ZM, that combines the merits of the Zernike moment (ZM) invariance and the continuous wavelet transform (CWT) time-frequency analysis. With the help of a wavelet time-frequency analysis, the THz pulse is expanded into a two-dimensional (2D) time-frequency plane that provides richer and more direct characteristic information in the time and frequency domain simultaneously. In addition, Zernike moments provide linearly independent descriptors for the 2D time-frequency intensity image and are invariant to THz signal affine transformations, such as peak shifting, baseline drifting, and scaling. In this manner, we obtain a set of features that exhibit a high capability to capture the concentrations of the target compounds and a high invariance of the different measuring instruments and the variable environment. This approach results in a more robust regression system with improved generalization properties with respect to standard methods. Experiments were then conducted on a THz dataset of pharmaceutical tablets acquired by two different THz instruments, and these confirmed the effectiveness of the proposed approach. Furthermore, CWT-ZM is an extensible framework that can be combined with various spectral qualitative and quantitative analysis algorithms.
The terahertz region presents promising opportunities for quantitative gas sensing in practical environments thanks to its access to molecular fingerprints and immunity to particulate scattering. With Terahertz time-domain spectroscopy (THz-TDS), molecular absorption and dispersion spectra can be rendered using femtosecond-laser-based interferometric system. However, procurement of accurate spectroscopic parameters in this uncharted spectral region and quantitative reduction of the broadband data remain challenging. We propose and validate two end-to-end machine learning models based on Gaussian process regression (GPR) for efficient THz-TDS sensing in both time- and frequency-domain (TD and FD). Quantitative CO sensing is accomplished for proof-of-concept demonstration of the method. Both TD- and FD-GPR models demonstrate accurate predictions in the presence of measurement noise and ambient interference, and the root-mean-square errors are below 0.6% with both simulated and experimental data. It is concluded that while the FD model benefits from immunity to dispersion distortions, the TD model noses out for its baseline-free nature and better elimination of temporally introduced noises. This work presents the first experimental demonstration of THz-TDS for quantitative gas sensing exploiting both TD and FD signals, showing promising prospects for real-time, multi-species, multi-parameter sensing under challenging scenarios such as chemically reactive or open-path atmospheric environments.
Concealed objects detection in terahertz imaging is an urgent need for public security and counter-terrorism. So far, there is no public terahertz imaging dataset for the evaluation of objects detection algorithms. This paper provides a public dataset for evaluating multi-object detection algorithms in active terahertz imaging. Due to high sample similarity and poor imaging quality, object detection on this dataset is much more difficult than on those commonly used public object detection datasets in the computer vision field. Since the traditional hard example mining approach is designed based on the two-stage detector and cannot be directly applied to the one-stage detector, this paper designs an image-based Hard Example Mining (HEM) scheme based on RetinaNet. Several state-of-the-art detectors, including YOLOv3, YOLOv4, FRCN-OHEM, and RetinaNet, are evaluated on this dataset. Experimental results show that the RetinaNet achieves the best mAP and HEM further enhances the performance of the model. The parameters affecting the detection metrics of individual images are summarized and analyzed in the experiments.
This paper proposes a supervised multinomial Bayesian learning algorithm for breast cancer detection using terahertz (THz) imaging of freshly excised murine tumors. The proposed algorithm utilizes a multinomial Bayesian probit regression approach, which establishes the link between THz data and classification results by using two different models, a polynomial regression model and a kernel regression model. Such a model-based learning approach employs only a small number of model parameters, thus it requires much less training data when compared with alternative deep learning methods. The training phase of the algorithm is performed by using the histopathology results of formalin-fixed, paraffin embedded (FFPE) samples as ground truth. There is usually a considerable shape mismatch between the freshly excised sample and its FFPE counterpart due to sample dehydration, and such mismatch negatively impacts the quality of the training data. We propose to address this challenge by using an innovative reliability-based training data selection method, where the reliability of the training data is quantified and estimated by using an unsupervised expectation maximization (EM) classification algorithm with soft probabilistic output. Experiment results demonstrate that the proposed multinomial Bayesian probit regression models with reliability-based training data selection achieve better performance than existing methods. Overall, these results demonstrate that the proposed supervised segmentation models represent a promising technique for the region detection with THz imaging of freshly excised breast cancer samples.
Terahertz (THz) communications are celebrated as key enablers for converged localization and sensing in future sixth-generation (6G) wireless communication systems and beyond. Instead of being a byproduct of the communication system, localization in 6G is indispensable for location-aware communications. Towards this end, we aim to identify the prospects, challenges, and requirements of THz localization techniques. We first review the history and trends of localization methods and discuss their objectives, constraints, and applications in contemporary communication systems. We then detail the latest advances in THz communications and introduce THz-specific channel and system models. Afterward, we formulate THz-band localization as a 3D position/orientation estimation problem, detailing geometry-based localization techniques and describing potential THz localization and sensing extensions. We further formulate the offline design and online optimization of THz localization systems, provide numerical simulation results, and conclude by providing lessons learned and future research directions. Preliminary results illustrate that under the same transmission power and array footprint, THz-based localization outperforms millimeter wave-based localization. In other words, the same level of localization performance can be achieved at THz-band with less transmission power or a smaller footprint.
Wireless communication at the terahertz (THz) frequency bands (0.1-10THz) is viewed as one of the cornerstones of tomorrow’s 6G wireless systems. Owing to the large amount of available bandwidth, if properly deployed, THz frequencies can potentially provide significant wireless capacity performance gains and enable high-resolution environment sensing. However, operating a wireless system at high-frequency bands such as THz is limited by a highly uncertain and dynamic channel. Effectively, these channel limitations lead to unreliable intermittent links as a result of an inherently short communication range, and a high susceptibility to blockage and molecular absorption. Consequently, such impediments could disrupt the THz band’s promise of high-rate communications and high-resolution sensing capabilities. In this context, this paper panoramically examines the steps needed to efficiently and reliably deploy and operate next-generation THz wireless systems that will synergistically support a fellowship of communication and sensing services. For this purpose, we first set the stage by describing the fundamentals of the THz frequency band. Based on these fundamentals, we characterize and comprehensively investigate seven unique defining features of THz wireless systems: 1) Quasi-opticality of the band, 2) THz-tailored wireless architectures, 3) Synergy with lower frequency bands, 4) Joint sensing and communication systems, 5) PHY-layer procedures, 6) Spectrum access techniques, and 7) Real-time network optimization. These seven defining features allow us to shed light on how to re-engineer wireless systems as we know them today so as to make them ready to support THz bands and their unique environments. On the one hand, THz systems benefit from their quasi-opticality and can turn every communication challenge into a sensing opportunity, thus contributing to a new generation of versatile wireless systems that can perform multiple functions beyond basic communications. On the other hand, THz systems can capitalize on the role of intelligent surfaces, lower frequency bands, and machine learning (ML) tools to guarantee a robust system performance. We conclude our exposition by presenting the key THz 6G use cases along with their associated major challenges and open problems. Ultimately, the goal of this article is to chart a forward-looking roadmap that exposes the necessary solutions and milestones for enabling THz frequencies to realize their potential as a game changer for next-generation wireless systems.
The problem of efficient ultra-massive multipleinput multiple-output (UM-MIMO) data detection in terahertz (THz)-band non-orthogonal multiple access (NOMA) systems is considered. We argue that the most common THz NOMA configuration is power-domain superposition coding over quasioptical doubly-massive MIMO channels. We propose spatial tuning techniques that modify antenna subarray arrangements to enhance channel conditions. Towards recovering the superposed data at the receiver side, we propose a family of data detectors based on low-complexity channel matrix puncturing, in which higher-order detectors are dynamically formed from lower-order component detectors. The proposed solutions are first detailed for the case of superposition coding of multiple streams in pointto-point THz MIMO links. Then, the study is extended to multiuser NOMA, in which randomly distributed users get grouped into narrow cell sectors and are allocated different power levels depending on their proximity to the base station. Successive interference cancellation is shown to be carried with minimal performance and complexity costs under spatial tuning. Approximate bit error rate (BER) equations are derived, and an architectural design is proposed to illustrate complexity reductions. Under typical THz conditions, channel puncturing introduces more than an order of magnitude reduction in BER at high signal-to-noise ratios while reducing complexity by approximately 90%.
With the vision to transform the current wireless network into a cyber-physical intelligent platform capable of supporting bandwidth-hungry and latency-constrained applications, both academia and industry turned their attention to the development of artificial intelligence (AI) enabled terahertz (THz) wireless networks. In this article, we list the applications of THz wireless systems in the beyond fifth-generation era and discuss their enabling technologies and fundamental challenges that can be formulated as AI problems. These problems are related to physical, medium/multiple access control, radio resource management, network, and transport layer. For each of them, we report the AI approaches, which have been recognized as possible solutions in the technical literature, emphasizing their principles and limitations. Finally, we provide an insightful discussion concerning research gaps and possible future directions.
Terahertz imaging and time-domain spectroscopy have been widely used to characterize the properties of test samples in various biomedical and engineering fields. Many of these tasks require the analysis of acquired terahertz signals to extract embedded information, which can be achieved using machine learning. Recently, machine learning techniques have developed rapidly, and many new learning models and learning algorithms have been investigated. Therefore, combined with state-of-the-art machine learning techniques, terahertz applications can be performed with high performance that cannot be achieved using modeling techniques that precede the machine learning era. In this review, we introduce the concept of machine learning and basic machine learning techniques and examine the methods for performance evaluation. We then summarize representative examples of terahertz imaging and time-domain spectroscopy that are conducted using machine learning.
Herein, we focus on convergent 6G communication, localization and sensing systems by identifying key technology enablers, discussing their underlying challenges, implementation issues, and recommending potential solutions. Moreover, we discuss exciting new opportunities for integrated localization and sensing applications, which will disrupt traditional design principles and revolutionize the way we live, interact with our environment, and do business. Regarding potential enabling technologies, 6G will continue to develop towards even higher frequency ranges, wider bandwidths, and massive antenna arrays. In turn, this will enable sensing solutions with very fine range, Doppler, and angular resolutions, as well as localization to cm-level degree of accuracy. Besides, new materials, device types, and reconfigurable surfaces will allow network operators to reshape and control the electromagnetic response of the environment. At the same time, machine learning and artificial intelligence will leverage the unprecedented availability of data and computing resources to tackle the biggest and hardest problems in wireless communication systems. As a result, 6G will be truly intelligent wireless systems that will provide not only ubiquitous communication but also empower high accuracy localization and high-resolution sensing services. They will become the catalyst for this revolution by bringing about a unique new set of features and service capabilities, where localization and sensing will coexist with communication, continuously sharing the available resources in time, frequency, and space. This work concludes by highlighting foundational research challenges, as well as implications and opportunities related to privacy, security, and trust.
Terahertz (THz)-band communications are celebrated as a key enabling technology for next-generation wireless systems that promises to integrate a wide range of data-demanding and delay-sensitive applications. Following recent advancements in optical, electronic, and plasmonic transceiver design, integrated, adaptive, and efficient THz systems are no longer far-fetched. In this article, we present a progressive vision of how the traditional "THz gap" will transform into a "THz rush" over the next few years. We posit that the breakthrough the THz band will introduce will not be solely driven by achievable high data rates, but more profoundly by the interaction between THz sensing, imaging, and localization applications. We first detail the peculiarities of each of these applications at the THz band. Then we illustrate how their coalescence results in enhanced environment-aware system performance in beyond-5G use cases. We further discuss the implementation aspects of this merging of applications in the context of shared and dedicated resource allocation, highlighting the role of machine learning.
Given the condition that protein conformation and activity are highly susceptible to environment factors such as temperature and pH, evaluation of protein conformation and activity is urgently needed in many fields. For example, most protein drugs need a stable and proper environment during production, storage and transportation, and it’s an enormous challenge to maintain protein activity throughout the whole process. Therefore, it’s necessary to ensure the safety and effectiveness of protein drugs by monitoring their activity before use. In our study, we presented an improved method for non-destructive evaluation of protein conformation and biological activity by terahertz spectroscopy combined with t-SNE-XGBoost. Firstly, bovine serum albumin (BSA) samples heated to different temperature were measured with THz-TDS. The obtained results indicated that native-conformation BSA will undergo transient states in the process of temperature induced denaturation. However, for any single given sample, it’s difficult to identify its conformation and activity directly by using the measured raw terahertz data. Therefore, we applied several different algorithms to the raw data for recognition of BSA samples with different conformation and activity induced by temperature. Finally, the models obtained by different algorithms were evaluated by calculating the root mean standard error of prediction (RMSEP) and the correlation coefficient of prediction (\(R_p\)). The THz-TDS plus t-SNE-XGBoost proved to be an effective non-destructive and label-free method for evaluation of protein conformation and activity. It can provide a new technique in many applications, such as pharmaceutical industry, clinical diagnosis and quality control.
Terahertz spectroscopy has proven to be a powerful tool for the study of condensed phase materials, opening research directions in a number of fields ranging from the pharmaceutical to semiconducting industries. Recent developments in terahertz technology have made this technique more accessible than ever before, and an increasing number of researchers are turning to terahertz spectroscopy for analysis and characterization of advanced materials. However, unlike mid-IR techniques, there do not exist any functional group specific transitions at terahertz frequencies, making the interpretation and assignment of terahertz spectral data more complex than complementary techniques. Through the aid of computational tools, incredible insights into the atomic-level dynamics occurring at terahertz frequencies have been uncovered, yet such highly accurate simulations require more care than traditional simulation methods in order to obtain such results. This review aims to highlight the recent advances in the computational assignment of terahertz spectral data, as well as showcasing common pitfalls to avoid, in order to demonstrate the utility of simulation methods for terahertz spectral assignment. Finally, cutting edge techniques and applications will be discussed, opening the door for future work in this exciting area of terahertz science.
Rice adulteration is a severe problem in agro-products and food regulatory agencies, suppliers, and consumers. In this study, to effectively distinguish whether high-quality rice is mixed with low-quality rice, detection and analysis of adulterated rice in five levels with different mixing proportions was conducted via terahertz spectroscopy and pattern recognition algorithms. Initially, samples were prepared and spectral data were acquired by using the terahertz transmission mode, and a principal component analysis (PCA) algorithm was applied to extract features from original spectrum information and reduce data dimensions. Subsequently, partial least squares-discriminant analysis (PLS-DA), support vector machine (SVM), and a back propagation neural network (BPNN) combined with the absorption spectra after different pretreatments, including standard normal variate (SNV) transformation, baseline correction (BC), and first derivative (1st derivative), were applied to establish the classification models. Results indicate that an SVM model employing the absorption spectra with a 1st derivative pretreatment exhibits the best discrimination ability, with an accuracy up to 97.33% in the prediction set. This result proves that terahertz spectroscopy combined with chemometric methods can be an effective tool to identify rice adulteration levels.
Investigations were initiated to develop terahertz (THz) techniques associated with machine learning methods of generalized regression neural network (GRNN) and back-propagation neural network (BPNN) to rapidly measure benzoic acid (BA) content in wheat flour. The absorption coefficient exhibited a maximum absorption peak at 1.94 THz, which generally increased with the content of BA additive. THz spectra were transformed into orthogonal principal component analysis (PCA) scores as the input vectors of GRNN and BPNN models. The best GRNN model was achieved with three PCA scores and spread value of 0.2. Compared with the BPNN model, GRNN model to powder samples could be considered very successful for quality control of wheat flour with a correlation coefficient of prediction (rp) of 0.85 and root mean square error of prediction of 0.10%. The results suggest that THz technique association with GRNN has a significant potential to quantitatively analyse BA additive in wheat flour.
Terahertz (THz) technique, a recently developed spectral method, has been researched and used for the rapid discrimination and measurements of food compositions due to its low-energy and non-ionizing characteristics. In this study, THz spectroscopy combined with chemometrics has been utilized for qualitative and quantitative analysis of myricetin, quercetin, and kaempferol with concentrations of 0.025, 0.05, and 0.1 mg/mL. The qualitative discrimination was achieved by KNN, ELM, and RF models with the spectra pre-treatments. An excellent discrimination (100% CCR in the prediction set) could be achieved using the RF model. Furthermore, the quantitative analyses were performed by partial least square regression (PLSR) and least squares support vector machine (LS-SVM). Comparing to the PLSR models, the LS-SVM yielded better results with low RMSEP (0.0044, 0.0039, and 0.0048), higher Rp (0.9601, 0.9688, and 0.9359), and higher RPD (8.6272, 9.6333, and 7.9083) for myricetin, quercetin, and kaempferol, respectively. Our results demonstrate that THz spectroscopy technique is a powerful tool for identification of three flavonols with similar chemical structures and quantitative determination of their concentrations.
This paper investigates terahertz (THz) imaging and classification of freshly excised murine xenograft breast cancer tumors. These tumors are grown via injection of E0771 breast adenocarcinoma cells into the flank of mice maintained on high-fat diet. Within 1 h of excision, the tumor and adjacent tissues are imaged using a pulsed THz system in the reflection mode. The THz images are classified using a statistical Bayesian mixture model with unsupervised and supervised approaches. Correlation with digitized pathology images is conducted using classification images assigned by a modal class decision rule. The corresponding receiver operating characteristic curves are obtained based on the classification results. A total of 13 tumor samples obtained from 9 tumors are investigated. The results show good correlation of THz images with pathology results in all samples of cancer and fat tissues. For tumor samples of cancer, fat, and muscle tissues, THz images show reasonable correlation with pathology where the primary challenge lies in the overlapping dielectric properties of cancer and muscle tissues. The use of a supervised regression approach shows improvement in the classification images although not consistently in all tissue regions. Advancing THz imaging of breast tumors from mice and the development of accurate statistical models will ultimately progress the technique for the assessment of human breast tumor margins.
Water vapor noise in the air affects the accuracy of optical parameters extracted from terahertz (THz) time-domain spectroscopy. In this paper, a numerical method was proposed to eliminate water vapor noise from the THz spectra. According to the Van Vleck–Weisskopf function and the linear absorption spectrum of water molecules in the HITRAN database, we simulated the water vapor absorption spectrum and real refractive index spectrum with a particular line width. The continuum effect of water vapor molecules was also considered. Theoretical transfer function of a different humidity was constructed through the theoretical calculation of the water vapor absorption coefficient and the real refractive index. The THz signal of the Lacidipine sample containing water vapor background noise in the continuous frequency domain of 0.5–1.8 THz was denoised by use of the method. The results show that the optical parameters extracted from the denoised signal are closer to the optical parameters in the dry nitrogen environment.
This paper describes the contents of the 2016 edition of the HITRAN molecular spectroscopic compilation. The new edition replaces the previous HITRAN edition of 2012 and its updates during the intervening years. The HITRAN molecular absorption compilation is composed of five major components: the traditional line-by-line spectroscopic parameters required for high-resolution radiative-transfer codes, infrared absorption cross-sections for molecules not yet amenable to representation in a line-by-line form, collision-induced absorption data, aerosol indices of refraction, and general tables such as partition sums that apply globally to the data. The new HITRAN is greatly extended in terms of accuracy, spectral coverage, additional absorption phenomena, added line-shape formalisms, and validity. Moreover, molecules, isotopologues, and perturbing gases have been added that address the issues of atmospheres beyond the Earth. Of considerable note, experimental IR cross-sections for almost 300 additional molecules important in different areas of atmospheric science have been added to the database. The compilation can be accessed through www.hitran.org. Most of the HITRAN data have now been cast into an underlying relational database structure that offers many advantages over the long-standing sequential text-based structure. The new structure empowers the user in many ways. It enables the incorporation of an extended set of fundamental parameters per transition, sophisticated line-shape formalisms, easy user-defined output formats, and very convenient searching, filtering, and plotting of data. A powerful application programming interface making use of structured query language (SQL) features for higher-level applications of HITRAN is also provided.
Terahertz spectroscopy has been investigated as a quick and non-destructive evaluation method to identify adulterated dairy products. Compared with the traditional identification methods, terahertz spectroscopy measurements can easily distinguish different adulterated dairy products according to the fatty acids, without pretreatment of the sample. Terahertz spectra are collected from samples of whole samples both without pretreatment. The difference between the spectra of samples can be observed with the fatty acids. This paper is using terahertz spectroscopy combination with chemometric tools to identify samples with PCA and SVM-DA. All samples can be correctly identified by SVM-DA models. These results demonstrate the performance of terahertz spectroscopy couple with chemometrics methods to identify adulterated food.
The combustion characteristics of fuel oils are closely related to both engine efficiency and pollutant emissions, and the analysis of oils and their additives is thus important. These oils and additives have been found to generate distinct responses to terahertz (THz) radiation as the result of various molecular vibrational modes. In the present work, THz spectroscopy was employed to identify a number of oils, including lubricants, gasoline and diesel, with different additives. The identities of dozens of these oils could be readily established using statistical models based on principal component analysis. The THz spectra of gasoline, diesel, sulfur and methyl methacrylate (MMA) were acquired and linear fittings were obtained. By using chemometric methods, including back propagation, artificial neural network and support vector machine techniques, typical concentrations of sulfur in gasoline (ppm-grade) could be detected, together with MMA in diesel below 0.5%. The absorption characteristics of the oil additives were also assessed using 2D correlation spectroscopy, and several hidden absorption peaks were discovered. The technique discussed herein should provide a useful new means of analyzing fuel oils with various additives and impurities in a non-destructive manner and therefore will be of benefit to the field of chemical detection and identification.
Terahertz (THz) spectroscopy is a promising method for analysing polar gas molecules mixed with unwanted aerosols due to its ability to obtain spectral fingerprints of rotational transition and immunity to aerosol scattering. In this article, dynamic THz spectroscopy of acetonitrile (CH3CN) gas was performed in the presence of smoke under the atmospheric pressure using a fibre-based, asynchronous-optical-sampling THz time-domain spectrometer. To match THz spectral signatures of gas molecules at atmospheric pressure, the spectral resolution was optimized to 1 GHz with a measurement rate of 1 Hz. The spectral overlapping of closely packed absorption lines significantly boosted the detection limit to 200 ppm when considering all the spectral contributions of the numerous absorption lines from 0.2 THz to 1 THz. Temporal changes of the CH3CN gas concentration were monitored under the smoky condition at the atmospheric pressure during volatilization of CH3CN droplets and the following diffusion of the volatilized CH3CN gas without the influence of scattering or absorption by the smoke. This system will be a powerful tool for real-time monitoring of target gases in practical applications of gas analysis in the atmospheric pressure, such as combustion processes or fire accident.
Spectroscopic techniques combined with chemometrics methods have proven to be effective tools for the discrimination of objects with similar properties. In this work, terahertz time-domain spectroscopy (THz-TDS) combined with discriminate analysis (DA) and principal component analysis (PCA) with derivative pretreatments was performed to differentiate transgenic rice (Hua Hui 1, containing the Cry1Ab protein) from its parent (Ming Hui 63). Both rice samples and the Cry1Ab protein were ground and pressed into pellets for terahertz (THz) measurements. The resulting time-domain spectra were transformed into frequency-domain spectra, and then, the transmittances of the rice and Cry1Ab protein were calculated. By applying the first derivative of the THz spectra in conjunction with the DA model, the discrimination of transgenic from non-transgenic rice was possible with accuracies up to 89.4% and 85.0% for the calibration set and validation set, respectively. The results indicated that THz spectroscopic techniques and chemometrics methods could be new feasible ways to differentiate transgenic rice.
The terahertz regime has particular value for liquid and biomolecular spectroscopy. In the case of liquids, terahertz is sensitive to relaxational and collective motions in liquids [1–13]. Applications include determination of sugar, alcohol, and water content. While there are no narrow band identification features for liquids in the terahertz range, the ability of THz to transmit through packaging materials and high sensitivity of relative water content is considered highly appealing for its use as a method to rapidly verify labeled contents. The determination of the water, sucrose, alcohol, liquid fuel, and petroleum content using terahertz have been demonstrated [1, 10]. The fundamental findings from terahertz measurements of liquids include the hydration number associated with solutes [14, 15], the extent of the perturbation of the liquid structure by the solute [16, 17], and the role of interactions in binary liquids [13, 18] . New collective mode vibrations have been identified for alcohols [19, 20], and the changes in the relaxational dynamics due to mixing, and the role of collective vibrations in ionic liquids [21–24]. In order to achieve these many findings, sensitive measurement techniques and data analysis have been developed. In parallel, great strides in modeling have been made to effectively model the picosecond dielectric response for these highly complex systems.
Nowadays terahertz spectroscopy is a well-established technique and recent progresses in technology demonstrated that this new technique is useful for both fundamental research and industrial applications. Varieties of applications such as imaging, non destructive testing, quality control are about to be transferred to industry supported by permanent improvements from basic research. Since chemometrics is today routinely applied to IR spectroscopy, we discuss in this paper the advantages of using chemometrics in the framework of terahertz spectroscopy. Different analytical procedures are illustrates. We conclude that advanced data processing is the key point to validate routine terahertz spectroscopy as a new reliable analytical technique.
Following recent advancements in Terahertz (THz) technology, THz communications are currently being celebrated as key enablers for various applications in future generations of communication networks. While typical communication use cases are over medium-range air interfaces, the inherently small beamwidths and transceiver footprints at THz frequencies support nano-communication paradigms. In particular, the use of the THz band for in-body and on-body communications has been gaining attention recently. By exploiting the accurate THz sensing and imaging capabilities, body-centric THz biomedical applications can transcend the limitations of molecular, acoustic, and radio-frequency solutions. In this paper, we study the use of the THz band for body-centric networks, by surveying works on THz device technologies, channel and noise modeling, modulation schemes, and networking topologies. We also promote THz sensing and imaging applications in the healthcare sector, especially for detecting zootonic viruses such as Coronavirus. We present several open research problems for body-centric THz networks.
Terahertz (THz)-band communications are a key enabler for future-generation wireless communication systems that promise to integrate a wide range of data-demanding applications. Recent advances in photonic, electronic, and plasmonic technologies are closing the gap in THz transceiver design. Consequently, prospect THz signal generation, modulation, and radiation methods are converging, and corresponding channel model, noise, and hardware-impairment notions are emerging. Such progress establishes a foundation for well-grounded research into THz-specific signal processing techniques for wireless communications. This tutorial overviews these techniques, emphasizing ultramassive multiple-input-multiple-output (UM-MIMO) systems and reconfigurable intelligent surfaces, vital for overcoming the distance problem at very high frequencies. We focus on the classical problems of waveform design and modulation, beamforming and precoding, index modulation, channel estimation, channel coding, and data detection. We also motivate signal processing techniques for THz sensing and localization.
We consider the problem of efficient blind parameter estimation in terahertz (THz)-band ultra-massive multiple-input multiple-output (UM-MIMO) systems. UM-MIMO antenna arrays are crucial to overcoming the distance problem in THz communications. Following recent advancements in THz transceiver design, such arrays are intrinsically compact and reconfigurable. We propose dynamically tuning the modulation types, operating frequencies, and indices of transmitting antenna elements for efficient resource utilization in THz UM-MIMO systems. Furthermore, we propose THz-specific signal processing techniques at the receiver side for blindly detecting the transmission parameters. In particular, we propose a low-complexity antenna index, frequency index, modulation mode, and modulation type detectors, and we study the complexity and performance tradeoffs of the proposed schemes.
Terahertz (THz)-band communications are currently being celebrated as a key technology that could fulfill the increasing demands for wireless data traffic in 6G wireless communications. Many challenges, such as high propagation losses and power limitations, which result in short communication distances, have yet to be addressed for this technology to be realized. Ultramassive multiple-input, multiple-output (UMMIMO) antenna systems have emerged as practical means for combatting this distance problem, thereby increasing system capacity. Toward that end, graphene-based nanoantennas have recently been proposed, as they can be individually tuned and collectively controlled in compact UM-MIMO array-of-subarrays (AoSA) architectures. In this article, we present a holistic overview of THz UM-MIMO systems. We assess recent advancements in transceiver design and channel modeling and discuss the major challenges and shortcomings of such designs by deriving the relationships among communication range, array dimensions, and system performance. We further highlight several research advances that could enhance resource allocation at the THz band, including waveform designs, multicarrier configurations, and spatial modulations (SMs). Based on this discussion, we highlight prospective use cases that can bring THz UM-MIMO into reality in the context of sensing, data centers, cell-free systems, and mid-range wireless communications.
In this work, a new approach to efficient gas radiation characteristics calculating is proposed to satisfy the demand for high accuracy and efficient calculation in many applications. This approach establishes a mapping relationship between gas condition parameters and radiation characteristics using a back-propagation neural network (BPNN). The line by line (LBL) model is utilized for the generation of training samples in the BPNN model. The values of pressure, temperature and component concentration are taken as input, and absorption coefficient values are taken as output. A case study of CO2 transmittance at 2250 - 2350 cm- 1 band is presented. The comparison and analysis of the results indicated that the BPNN model has a high accuracy of LBL fitting and is insensitive to the input. Although the training time of BPNN is long, once the training is completed, the computational efficiency is very high. Compared to the look-up table method or other accelerated methods using parameter pre-calculation, the BPNN method occupies much less storage space. It can replace the LBL model to a certain extent when dealing with the needs of high precision and high-speed computing.
A novel approach was presented to characterize microstructural features of thermal barrier coatings (TBCs) using terahertz spectroscopy based on machine learning algorithms. In this study, the microstructures of yttria-stabilized zirconia (YSZ) atmospheric-plasma-sprayed (APS) thermal barrier coatings were regulated by choosing different kinds of spray powders, distances and power during processing. A terahertz time-domain spectroscopy system configurated transmission mode with an incidence angle of 0° was employed to estimate terahertz properties of porous YSZ ceramic coatings, including refractive index, extinction coefficient and relative time-domain broadening ratio. The variation tendency of terahertz properties of YSZ ceramic coatings with different microstructure features (porosity, pore to crack ratio, pore size) were investigated. Principal component analysis (PCA) method was adopted to reduce the dimensions of refractive index and extinction coefficient spectra data at the range of 0.6–1.4 THz and to ensure that different terahertz properties could be treated as inputs with similar weights during modeling. Three models (multiple linear regression (MLR), back-propagation (BP) neural network and support vector machine (SVM)) were set up to conduct regression analysis. As a result, according to the contribution rates of eigenvectors, the top one principal component of refractive index spectra data and the top two principal components of extinction coefficient spectra data were selected as the model inputs. The correlation coefficient comparisons showed that the characterization accuracy of PCA-SVM reached by over 95% and outperformed the other models. Finally, this study proposed that THz nondestructive technology combined with machine learning technique is efficient and feasible for microstructural features characterization and has profound implications for the structure integrity of TBCs evaluation in gas turbine blades.
With 5G Phase 1 finalized and 5G Phase 2 recently defined by 3GPP, the mobile communication community is on the verge of deciding what will be the Beyond-5G (B5G) system. B5G is expected to further enhance network performance, for example, by supporting throughput per device up to terabits per second and increasing the frequency range of usable spectral bands significantly. In fact, one of the main pillars of 5G networks has been radio access extension to the millimeter-wave bands. However, new envisioned services, asking for more and more throughput, require the availability of one order of magnitude more spectrum chunks, thus suggesting moving the operations into the THz domain. This move will introduce significant new multidisciplinary research challenges emerging throughout the wireless communication protocol stacks, including the way the mobile network is modeled and deployed. This article, therefore, provides a brief survey of the challenges and opportunities of THz band operation in wireless communication, along with some potential applications and future research directions.
Future smart vehicles will incorporate high-data-rate communications and high-resolution radar sensing capabilities operating in the millimeter-wave and higher frequencies. These two systems are preparing to share and reuse many common functionalities, such as steerable millimeter-wave antenna arrays. Motivated by this growing overlap, which is advanced further by space and cost constraints, the vehicular community is pursuing a vision of unified vehicular communications and radar sensing that represents a major paradigm shift for next-generation connected and self-driving cars. This article outlines a path to materialize this decisive transformation. We begin by reviewing the latest developments in hybrid vehicular communications and radar systems, and then propose a concept of unified channel access over millimeter-wave and higher frequencies. Our supporting system-level performance characterization relies upon real-life measurements and extensive ray-based modeling to confirm the significant improvements brought by our proposal to mitigating the interference and deafness effects. Since our results aim to open the door to unified vehicular communications and radar sensing, we conclude by outlining the potential research directions in this rapidly developing field.
The automatic extraction of the targets which we are interested from a given image is the fundamental of the automatic detection and identification for security screening systems based on imaging technologies. Suffering from the comparatively low signal to noise ratio (SNR), the automatic detection of targets in a passive Terahertz (THz) imaging system facing great challenges, but in urgent necessary. In this paper, a comprehensive method for automatic detection of concealed targets in passive THz image by making the best use of the ‘block statistics uniformity’ properties of the passive images is firstly studied. A theoretical model for the ‘featured regions’ decomposition based on the minimization of a ‘fit energy’ functional with respect to a ‘surface function’ is established, to overcome the drawbacks of conventional methods with gradient-based edge operators for their unsuccessful application in low SNR passive images with blurred boundaries. Based on above theoretical basis and taking advantages of the distinguished contrasts of the convergent ‘surface function’ in different ‘featured regions’, an automatic detection algorithm with three steps was further developed to automatically extract the number, the locations and the shapes of all the concealed targets, with the shape of each target derived as the contour point series arranged in clockwise direction. With plenty of experimental results in 0.2THz band, it's found that, the proposed method has high detection accuracy about 95% with quite good real-time performance, even for the single channel proof-of-state system with low SNR. The theorem, algorithm and results in this paper may have important applications in unmanned and intelligent security screening systems without any artificial interventions.
The field of terahertz integrated technology has undergone significant development in the past ten years. This has included work on different substrate technologies such as III–V semiconductors and silicon, work on field-effect transistor devices and heterojunction bipolar devices, and work on both fully electronic and hybrid electronic–photonic systems. While approaches in electronic and photonics can often seem distinct, techniques have blended in the terahertz frequency range and many emerging systems can be classified as photonics-inspired or hybrid. Here, we review the development of terahertz integrated electronic and hybrid electronic–photonic systems, examining, in particular, advances that deliver important functionalities for applications in communication, sensing and imaging. Many of the advances in integrated systems have emerged, not from improvements in single devices, but rather from new architectures that are multifunctional and reconfigurable and break the trade-offs of classical approaches to electronic system design. We thus focus on these approaches to capture the diversity of techniques and methodologies in the field. © 2018, The Author(s), under exclusive licence to Springer Nature Limited.
The ability to diagnose oral lichen planus (OLP) based on saliva analysis using THz time-domain spectroscopy and chemometrics is discussed. The study involved 30 patients (2 male and 28 female) with OLP. This group consisted of two subgroups with the erosive form of OLP (n = 15) and with the reticular and papular forms of OLP (n = 15). The control group consisted of six healthy volunteers (one male and five females) without inflammation in the mucous membrane in the oral cavity and without periodontitis. Principal component analysis was used to reveal informative features in the experimental data. The one-versus-one multiclass classifier using support vector machine binary classifiers was used. The two-stage classification approach using several absorption spectra scans for an individual saliva sample provided 100% accuracy of differential classification between OLP subgroups and control group.
Discrimination of geographical origin of extra-virgin olive oils (EVOOs) is of great importance for legislation and consumers worldwide. The feasibility of a rapid discrimination of four different geographical origins of EVOOs with terahertz spectroscopy system was examined. Different chemometrics including least squares-support vector machines (LS-SVM), back propagation neural network (BPNN) and random forest (RF) combined with principal component analysis (PCA), genetic algorithm (GA) were compared to obtain the best discrimination model. The results demonstrated that there were apparent differences among the four different geographical origins of EVOOs in fatty acid compositions and the absorbance spectra, and an excellent classification (accuracy was 96.25% in prediction set) could be achieved using the LS-SVM method combine with GA. It can be concluded that THz spectroscopy together with chemometrics would be a promising technique to rapid discriminate the geographical origin of EVOOs with high efficiency.
Purpose
This paper aims to provide a technical insight into a selection of recent developments and applications involving terahertz sensing technology.
Design/methodology/approach
Following an introduction, the first part of this paper considers a selection of research activities involving terahertz radiation sources and detectors. The second part seeks to illustrate how the technology is exerting a commercial impact and discusses a number of product developments and applications.
Findings
Terahertz sensing is a rapidly developing field and a strong body of research seeks to develop sources and detectors with enhanced features which often exploit novel materials, phenomena and technologies. Commercialisation is gathering pace, and a growing number of companies are producing terahertz sensing and imaging products which are finding a diversity of applications.
Originality/value
This provides details of recent research, product developments and applications involving terahertz sensing technology.
Aflatoxins contaminate and colonize agricultural products, such as grain, and thereby potentially cause human liver carcinoma. Detection via conventional methods has proven to be time-consuming and complex. In this paper, the terahertz (THz) spectra of aflatoxin B1 in acetonitrile solutions with concentration ranges of 1-50 μg/ml and 1-50 μg/l are obtained and analyzed for the frequency range of 0.4-1.6 THz. Linear and nonlinear regression models are constructed to relate the absorption spectra and the concentrations of 160 samples using the partial least squares (PLS), principal component regression (PCR), support vector machine (SVM), and PCA-SVM methods. Our results indicate that PLS and PCR models are more accurate for the concentration range of 1-50 μg/ml, whereas SVM and PCA-SVM are more accurate for the concentration range of 1-50 μg/l. Furthermore, ten unknown concentration samples extracted from mildewed maize are analyzed quantitatively using these methods.
Petrochemicals, one of the most important energy sources, contribute to the remarkable development of human civilization. Therefore, the development of a kind of fast, safe, reliable and nondestructive detection technology is essential. Terahertz (THz) spectroscopy, containing abundant physical, chemical, and structural information of materials, shows significant applications in the fields of physics, chemistry, materials science, medicine, pharmacy and biology. As a promising detection technology, THz technology provides a new reliable analytic method in liquid petrochemicals detection due to the fact that low-frequency vibrational and rotational motions of hydrocarbon molecules lie in the terahertz region. In this paper, we review the applications of the liquid petrochemicals detection based on the terahertz time-domain spectroscopy (THz-TDS) system, mainly containing the analysis of molecular properties, qualitative identification, quantitative analysis and the terahertz metamaterials sensing. In addition, we propose the further exploration of terahertz technology in the field of petrochemical industry.
The application of pattern recognition methodology within chemistry, biology and other science domains, especially in security systems is becoming more and more important. Many classification algorithms are available in literature but decision trees are the most commonly exploited because of their ease of implementation and understanding in comparison to other classification algorithms. Decision trees are powerful and popular tools for classification and prediction. In contrast to neural networks, decision trees represent rules, which can readily be expressed so that humans can understand them or even directly use in a database. In this paper we present an algorithm of construction of decision trees and a classification rule extraction based on a logical relationship between attributes and a generalized decision function. Moreover, correctness and efficiency of the algorithm was experimentally validated in a terahertz system, where spectra of explosives were measured in reflection configuration.
Nanotechnologies promise new solutions for several applications in the biomedical, industrial and military fields. At the nanoscale, a nanomachine is considered as the most basic functional unit which is able to perform very simple tasks. Communication among nanomachines will allow them to accomplish more complex functions in a distributed manner. In this paper, the state of the art in molecular electronics is reviewed to motivate the study of the Terahertz Band (0.1-10.0 THz) for electromagnetic (EM) communication among nano-devices. A new propagation model for EM communications in the Terahertz Band is developed based on radiative transfer theory and in light of molecular absorption. This model accounts for the total path loss and the molecular absorption noise that a wave in the Terahertz Band suffers when propagating over very short distances. Finally, the channel capacity of the Terahertz Band is investigated by using this model for different power allocation schemes, including a scheme based on the transmission of femtosecond-long pulses. The results show that for very short transmission distances, in the order of several tens of millimeters, the Terahertz channel supports very large bit-rates, up to few terabits per second, which enables a radically different communication paradigm for nanonetworks.
Demand for high levels of quality and safety in agricultural products and food requires appropriate analytical techniques for analysis both during and after production. Desirable techniques should be quick, easy, and safe to use; require minimal or no sample preparation; avoid sample destruction; and be accurate. In this study, the recent technical applications of terahertz spectroscopy to identify and classify, qualitatively and quantitatively analyze, evaluate, and safely control agricultural products and food are reviewed. The challenges and future outlook of terahertz spectroscopy are discussed.
Absorption and reflection spectra have successfully been used for substance identification, which is also applicable to the THz spectral range. For optical spectroscopy, the Kramers–Kronig transformation (KKT) is a powerful tool to determine the complex refractive index n˜ (with k, n absorption and refractive indices, respectively) from either an absorption or reflection measurement. By terahertz time-domain spectroscopy (THz-TDS) materials are probed with short pulses of radiation. The detection is sensitive to the sample on both the amplitude and the phase of the electrical field. Thus, THz-TDS spectroscopy can provide more information than conventional Fourier-transform spectroscopy, by which a power spectrum is measured. In the case of transmission measurements, for instance, formulae exist by which the frequency-dependent complex refractive index is directly calculated from the time-dependent electrical field (waveform) without the necessity of a KKT. In the case of reflection experiments, a comparable computation is possible for the frequency-dependent phase angle; here a KKT can help to recover essential parameters. We present a combination of KKT and TDS methods for the calculation of optical constants in the THz regime.
Terahertz time domain spectroscopy (THz-TDS) has a wide range of applications from semiconductor diagnostics to biosensing. Recent attention has focused on bio-applications and several groups have noted the ability of THz- TDS to differentiate basal cell carcinoma tissue from healthy dermal tissue ex vivo.1 The contrast mechanism is unclear but has been attributed to increased interstitial water in cancerous tissue. In this work we investigate the THz response of human osteosarcoma cells and normal human bone cells grown in culture to isolate the cells' responses from other effects. A classification algorithms based on a frequency selection by genetic algorithm is used to attempt to differentiate between the cell types based on the THz spectra. Encouraging preliminary results have been obtained.
Over the past three decades a new spectroscopic technique with unique possibilities has emerged. Based on coherent and time-resolved detection of the electric field of ultrashort radiation bursts in the far-infrared, this technique has become known as terahertz time-domain spectroscopy (THz-TDS). In this review article the authors describe the technique in its various implementations for static and time-resolved spectroscopy, and illustrate the performance of the technique with recent examples from solid-state physics and physical chemistry as well as aqueous chemistry. Examples from other fields of research, where THz spectroscopic techniques have proven to be useful research tools, and the potential for industrial applications of THz spectroscopic and imaging techniques are discussed.
Recent development of the terahertz time domain spectroscopy (THz-TDS) and its application to solids have been reviewed. This
spectroscopy is unique in that the time-domain wave forms are measured at first and the complex optical constants are deduced
directly by the Fourier transformation of them without resort to the Kramers-Kronig analysis. Various types of the THz-TDS
systems are briefly described. Applications of the THz-TDS to various solids, i.e., semiconductors, superconductors, polymers,
photonic crystals, and so on are also presented to demonstrate how widely this spectroscopy is applicable to characterization
of solids.
Echo state networks (ESNs) constitute a novel approach to recurrent neural network (RNN) training, with an RNN (the reservoir) being generated randomly, and only a readout being trained using a simple computationally efficient algorithm. ESNs have greatly facilitated the practical application of RNNs, outperforming classical approaches on a number of benchmark tasks. In this paper, we introduce a novel Bayesian approach toward ESNs, the echo state Gaussian process (ESGP). The ESGP combines the merits of ESNs and Gaussian processes to provide a more robust alternative to conventional reservoir computing networks while also offering a measure of confidence on the generated predictions (in the form of a predictive distribution). We exhibit the merits of our approach in a number of applications, considering both benchmark datasets and real-world applications, where we show that our method offers a significant enhancement in the dynamical data modeling capabilities of ESNs. Additionally, we also show that our method is orders of magnitude more computationally efficient compared to existing Gaussian process-based methods for dynamical data modeling, without compromises in the obtained predictive performance.
THz spectral database
- E Heilweil
- M Campbell