Figure - available via license: Creative Commons Attribution 2.0 Generic
Content may be subject to copyright.
Source publication
Surface electromyographic (S-EMG) signal processing has been emerging in the past few years due to its non-invasive assessment of muscle function and structure and because of the fast growing rate of digital technology which brings about new solutions and applications. Factors such as sampling rate, quantization word length, number of channels and...
Similar publications
Deep convolutional neural networks (CNNs) are used for image denoising via automatically mining accurate structure information. However, most of existing CNNs depend on enlarging depth of designed networks to obtain better denoising performance, which may cause training difficulty. In this paper, we propose a multi-stage image denoising CNN with th...
The future is mainly focused on image brightness and the capacity that required storing the image. The sharp images provide better in-formation than the blur images. To overcome from the blurriness in the image, we use image enhancement techniques. Image fusion used to overcome information loss in the image. This paper is provided with image enhanc...
This paper proposes a new tunable-Q wavelet transform (TQWT) based feature extraction for classification of power quality (PQ) disturbances. The filters are designed to extract the fundamental frequency component from the complete voltage/current signal. This is achieved by tuning the Q-factor (Q) and redundancy (r) of the wavelet by primarily inve...
In this paper a coupled dictionary learning mechanism with mapping function is proposed in the wavelet domain for the task of Single-Image Super-Resolution. Sparsity is used as the invariant feature for achieving super-resolution. Instead of using a single dictionary multiple compact dictionaries are proposed in the wavelet domain. Such dictionarie...
The need for image restoration is encountered in many practical applications. For instance, distortion due to Additive White Gaussian Noise (AWGN) or in some cases the multiplicative (speckle) one can be caused by poor quality image acquisition. Wavelet denoising attempts to remove these types of noise present in the signal while preserving the sig...
Citations
... Although this solution is widespread in research and clinical practice, the use of temporary electrodes together with bulky recording devices reduces patient comfort and makes it difficult to record the EMG signal for a prolonged period of time. Therefore, the need to develop innovative sEMG recording devices that can be embedded into wearable electronic systems, including processing capabilities (e.g., data compression [12]), is evident. Recent research has gone in the direction of designing wearable sEMG solutions, which include electrodes that are printed or integrated into flexible supports, which also include interconnects [13]. ...
Driven by the fast-expanding market, wearable technologies have rapidly evolved [...]
... Therefore, the quality of the replicated signal and compression factor are always in trade-off. Several methods are documented in the literature that are based on linear and non-linear transform like Discrete Wavelet Transforms (DWT) and Discrete Cosine Transforms (DCT) [4,5,10,15,18,21,23], modified DWT [13,15,17], swarm based [1], DPCM [3,4], Vector quantization [2], artificial neural network [23] and image encoders JPEG2000 and video encoder H.264/AVC [6,7,8,11,12,14,16,19,20]. Furthermore, researchers are striving to show the trade-off between computational complexity and distortions in reconstructed signals [22]. JPEG2000 encoder (used to compress the image signal) and H.264/AVC (video encoder) were tested by Costa et al. [7, 8, 11, and 12] to demonstrate that they may also be used to compress EMG signals and achieved a higher compression factor when compared to alternatives. ...
... JPEG2000 encoder (used to compress the image signal) and H.264/AVC (video encoder) were tested by Costa et al. [7, 8, 11, and 12] to demonstrate that they may also be used to compress EMG signals and achieved a higher compression factor when compared to alternatives. For the required outcomes, they additionally applied DWT with entropy encoding [5] and spectrum segmentation [18]. The Artificial Bee Colony (ABC) swarm algorithm is used by Hosney et al. [1] in 2018 to choose the ideal characteristics of bio signals like ECG, EMG and EEG, needed to preserve the reconstructed signal's quality for certain compression factor. ...
Modern technology allows for the development of improved telehealth monitor and care systems for patients who are in rural areas. Due to very effective biological signal data compression and reconstruction technologies, it is now conceivable. Several researchers have created and used a variety of EMG compression and reconstruction techniques over the last 20 years. The process of data compression involves lowering the number of bits required to present the data. It is done to reduce bandwidth of the network, free up storage space, and fast transfer of data while preserving passable signal quality. Performance indicators including the percentage root mean square difference (PRD) and Compression Factor (CF) were examined and compared for these approaches. High CF increases savings but lowers the signal quality that is compressed. Increased PRD increases the quality at the cost of CF. Hence, CF and PRD need to be in balanced. The review and analysis of 21 PRD and CF-based published papers is the main emphasis of this study. CF and signal reconstruction quality, which is assessed through PRD, are of utmost importance for an effective compression. The correlation among these two is taken into consideration because they are both dependent characteristics. Because it is a systematic and statistical procedure that combines the findings of earlier studies, offers a common answer, and identifies similarities and comparative behavior among many studies, meta-analysis is utilized for analysis. The aggregate correlation value of all 21 (To maintain data homogeneity, three studies were excluded) research works, is 0.77, and this number is used to categorize the works. Due of the analysis's 13.86% heterogeneity (I 2) result, all research papers fall under the same category. In [16], researchers utilized a sample size of 80 and employed a two-dimensional approach for compression, specifically using a recurrent pattern matching algorithm called the multi-dimensional multiscale parser (MMP). According to the statistical analysis in the study, de Melo WC (2016) achieved an average PRD of 6.40, indicating a close resemblance between the original and reconstructed EMG signals. Additionally, the study reported a higher CF of
... It is expressed as the moving average of the full-wave rectified EMG signal. This is defined as [7,8,9,10,11,12]: ...
... Simple Square Integral (SSI) -the Simple Square Integral (SSI) expresses the energy of the EMG signal as a useable feature [13,14]. This is defined as [7,8,9,10,11,12]: ...
... Variance of EMG (VAR) -The Variance of EMG (VAR) expresses the power [13,14] of the EMG signal as a useable feature. This is defined as [7,8,9,10,11,12]: ...
... With Lossy methods, however, significant compression can be achieved, at the expense of losing some information in the data. Previous studies on EMG data compression have applied widely used transform-based lossy compression techniques including discrete wavelet transform (DWT) coding with optimized basis function [9,10], embedded zero-trees of DWT coefficients [11,12], vector quantization of the discrete cosine transform (DCT) and DWT coefficients [13,14], dynamic bit allocation of the DWT coefficients [15,16], the multidimensional multiscale parser algorithm [17], a codebook linear prediction [18], the JPEG2000 and H.264/AVC-intra transforms [19], compressed sensing [20], the SPIHT and arithmetic coding [14], entropy coding [16], and video codec algorithm [21]. In contrast to the previous studies described above, this work proposes a fundamentally different strategy by applying a deep learning approach for EMG compression. ...
... With Lossy methods, however, significant compression can be achieved, at the expense of losing some information in the data. Previous studies on EMG data compression have applied widely used transform-based lossy compression techniques including discrete wavelet transform (DWT) coding with optimized basis function [9,10], embedded zero-trees of DWT coefficients [11,12], vector quantization of the discrete cosine transform (DCT) and DWT coefficients [13,14], dynamic bit allocation of the DWT coefficients [15,16], the multidimensional multiscale parser algorithm [17], a codebook linear prediction [18], the JPEG2000 and H.264/AVC-intra transforms [19], compressed sensing [20], the SPIHT and arithmetic coding [14], entropy coding [16], and video codec algorithm [21]. In contrast to the previous studies described above, this work proposes a fundamentally different strategy by applying a deep learning approach for EMG compression. ...
Efficient storage and transmission of electromyogram (EMG) data are important for emerging applications such as telemedicine and big data, as a vital tool for further advancement of the field. However, due to limitations in internet speed and hardware resources, transmission and storage of EMG data are challenging. As a solution, this work proposes a new method for EMG data compression using deep convolutional autoencoders (CAE). Eight-channel EMG data from 10 subjects, and high density EMG data from 18 subjects, were investigated for compression. The CAE architecture was designed to extract an abstract data representation that is heavily compressed, but from which the salient information for classification can be effectively reconstructed. The proposed method attained efficient compression; for CR=1600, the average PRDN (percentage RMS difference normalized) was 31.5% and the wrist motions classification accuracy (CA) reduced roughly 5%. The CAE substantially outperformed the state-of-the-art high-efficiency video coding and a well known wavelet-thresholding compression technique. Moreover, by reducing the bit-resolution of the CAEs compressed data from 24 bits to 6 bits, an additional 4-fold compression was achieved without significant degradation of the reconstruction performance. Furthermore, the CAEs inter subject performance was promising; e.g. for CR=1600, the PRDN for the inter-subject case was only 2.6% less than that of the within-subject performance. The powerful EMG compression performance with remarkable reconstruction results reflects the CAEs potential as an automatic end to end approach with the ability to learn the complete encoding and decoding process. Furthermore, the excellent inter-subject performance demonstrates the generalizability and usability of the proposed approach.
... The image compression techniques applied on surface EMG signals in [6,7]. Trabuco et al. [8] employed discrete wavelet transform (DWT) with four different spectral functions for bit allocation. Vector quantization with SPIHT coding and arithmetic coding applied for EMG signal compression [9]. ...
The chances of morbidity and mortality of premature babies increases every year around the world. Appropriate care of these babies at the right moment through telemedicine and ambulatory monitoring is vital. The size of the uterine physiological signals restrains the practical applications. A lossless compression method proposed in the study as a combination of discrete cosine transform (DCT) and Huffman coding. The DCT components below the 2 Hz frequency were computed. The quantized DCT coefficients were coded by a Huffman coder on the transmitter site. A zero set instead of the DCT coefficients above 2 Hz, added at the receiver site. Inverse DCT applied to get the reconstructed signal. Uterine Magnetomyography (MMG) signals downloaded from the Physionet database were used in this study. The results indicate that the proposed algorithm is best suitable for the lossless compression of MMG signals.
... All signals were sampled at 2 kHz and digitized with 2 bytes/sample. The duration of the signals varies from 3 to 6 minutes [17]. ...
... All signals were sampled at 2 kHz and quantized on 16 bits. The duration of the signals varies from 3 to 6 minutes [17]. ...
... In this approach, the EMG signals were transformed to a 2D matrix and then the DCT or DWT were applied to this 2D matrix to compute the coefficients. Recently, Trabuco et al. [40,41] proposed EMG compression algorithms based on DWT and dynamic bit allocation to code the wavelet coefficients. ...
In this paper, an algorithm is proposed for efficient compression of bio-signals based on discrete Tchebichef moments and Artificial Bee Colony (ABC). The Tchebichef moments are used to extract features of the bio-signals, then, the ABC algorithm is used to select of the optimum features which achieve the best bio-signal quality for a specific compression ratio (CR). The proposed algorithm has been tested by using different datasets of Electrocardiogram (ECG), Electroencephalogram (EEG), and Electromyogram (EMG). The optimum feature selection using ABC significantly improve the quality of the reconstructed bio-signals. Different numerical experiments are performed to compress different records of ECG, EEG and EMG bio-signals by using the proposed algorithm and the most recent existing methods. The performance of the proposed algorithm and the other existing methods are evaluated using different metrics such as CR, PRD, and peak signal to noise ratio (PSNR). The comparison has shown that, at the same CR, the proposed compression algorithm yields the best quality of the reconstructed signals over the other existing methods. © 2018 Nalecz Institute of Biocybernetics and Biomedical Engineering of the Polish Academy of Sciences
... Entretanto, para a exploração da correlação bidimensional no processo de compressão, conforme se descreve nesses trabalhos, há a necessidade de se ter o sinal eletromiográfico completo previamente adquirido. Em [12], os autores propõem um algoritmo de compressão utilizando decomposição em wavelets e alocação dinâmica de bits em sub-bandas. Em comum, todos esses trabalhos citados fazem uso de transformadas ortogonaise da representação esparsa obtida no domínio transformadono processo de compressão. ...
... For example, a vector quantization was applied on the transformed wavelet coefficient vector [35]. Another approach used mathematical models [36] or neural networks [37], [38] to approximate the spectral magnitude shape in wavelet domain. Consequently, dynamic bit allocation was carried out for quantization of wavelettransformed coefficients. ...
... Wavelet transform based encoders with dynamic bit allocation for the transformed coefficients quantization was presented in [36]. The transformed coefficient vector, in this proposal, is quantized according to the mathematical model used for spectral shape estimation. ...
... The two metrics are most commonly used in scientific literature to evaluate the performance of the encoders: The Compression Factor (CF) and the Percent Residual Difference (PRD) [27], [29], [30], [32], [36]- [41], [44], [48]. The compression where O s is the number of bits required for storing the original data, and C s is the number of bits to represent the compressed data. ...
This paper presents algorithms designed for one-dimensional (1-D) and 2-D surface electromyographic (S-EMG) signal compression. The 1-D approach is a wavelet transform based encoder applied to isometric and dynamic S-EMG signals. An adaptive estimation of the spectral shape is used to carry out dynamic bit allocation for vector quantization of transformed coefficients. Thus, an entropy coding is applied to minimize redundancy in quantized coefficient vector and to pack the data. In the 2-D approach algorithm, the isometric or dynamic S-EMG signal is properly segmented and arranged to build a 2-D representation. The high efficient video codec is used to encode the signal, using 16-bit-depth precision, all possible coding/prediction unit sizes, and all intra-coding modes. The encoders are evaluated with objective metrics, and a real signal data bank is used. Furthermore, performance comparisons are also shown in this paper, where the proposed methods have outperformed other efficient encoders reported in the literature.
... Due to the scarcity of research into the compression of EMG signals of maternal patients, studies of EMG signals, regardless of disease type, were used for comparison (refer to Section 4.2). The studies by Balouchestani [14], Itiki [15], Norris [16], Berger [17], Filho [18], and Trabuco [7] were selected. Specific summary details of the selected researches are shown in Table 1. ...
... Trabuco [7] Algorithm based on discrete wavelet transform for spectral decomposition and de-correlation. Does not support real-time. ...
An estimated 15 million babies are born prematurely every year worldwide, and suffer from disabilities. Appropriate care of these pre-term babies immediately after birth through telemedicine monitoring is vital. However, problems associated with a limited bandwidth and network overload due to the excessive size of the electromyography (EMG) signal impede the practical application of such medical information systems. Therefore, this research proposes an EMG uterine monitoring transmission solution (EUMTS), a lossless efficient real-time EMG transmission solution that solves such problems through efficient EMG data lossless compression. EMG data samples obtained from the Physionet PhysioBank database were used. Solution performance comparisons were conducted using Lempel-ZivWelch (LZW) and Huffman methods, in addition to related researches. The LZW and Huffman methods showed CRs of 1.87 and 1.90, respectively, compared to 3.61 for the proposed algorithm. This was relatively high compared to related researches, even when considering that those researches were lossy whereas the proposed research was lossless. The results also showed that the proposed algorithm contributes to a reduction in battery consumption by reducing the wake-up time by 1470.6 ms. Therefore, EUMTS will contribute to providing an efficient wireless transmission environment for the prediction of pre-term delivery, enabling immediate interventions by medical professionals. Another novel point of EUMTS is that it is a lossless algorithm, which will prevent any misjudgement by clinicians because the data will not be distorted. Pre-term babies may receive point-of-care immediately after birth, preventing exposure to the development of disabilities.
