Figure - available from: Computational and Mathematical Methods in Medicine
This content is subject to copyright. Terms and conditions apply.
Enlarged view of the marked green box in Figure 7. (a) A 2D slice view. (b) Hessian-based multiscale filtering. (c) Hessian-based multiscale filtering combined with Chan-Vese model. (d) The proposed method.
Source publication
Vascular segmentation plays an important role in medical image analysis. A novel technique for the automatic
extraction of vascular trees from 2D medical images is presented, which combines Hessian-based multiscale filtering and a modified level set method. In the proposed algorithm, the morphological top-hat transformation is firstly adopted to at...
Citations
... In order to assess their approach, authors generated synthetic images. 19 Deep fully convolutional networks (FCN) have recently been developed as a result of recent advancements in computer vision, which have improved the performance of semantic segmentation and enabled to outperform other competitors in the field of medical imaging. [20][21][22] Convolutional neural networks (CNNs), a type of artificial neural network (ANN) designed for image processing, have emerged as the method of choice for vessel segmentation, including the hepatic vasculature, and have shown encouraging results in recent applications. ...
... The primary idea is that the images are convolved using 3D Gaussian filters at various scales, and the shape of the local structures within the images is determined by analyzing the eigenvalues of the Hessian matrix at each pixel or voxel in terms of a response function. 19 To determine the maximum vesselness values, the Hessian enhancement is applied to the target image at various scales on each voxel. The contrast of the small vessels will be improved by the Hessian enhancement at small scale values (1-3), whereas the vessels with large radii will be enhanced at big scale values (4−8). ...
... For several reasons, it is difficult to do quantitative comparisons with other segmentation approaches described in the literature. 13,19,31,[40][41][42][43][44][45][46] First, there are not many studies on this topic of liver vasculature segmentation. Some authors used their own clinical data sets, others employed artificial data sets, and still others confirmed their findings using images of created forms (phantoms) resembling veins. ...
Purpose:
Liver hepatic vessels segmentation is a crucial step for the diagnosis process in patients with hepatic diseases. Segmentation of liver vessels helps to study the liver internal segmental anatomy that helps in the preoperative planning of surgical treatment.
Methods:
Recently, the convolutional neural networks (CNN) have been proved to be efficient for the task of medical image segmentation. The paper proposes an automatic deep learning-based system for liver hepatic vessels segmentation of Computed Tomography (CT) datasets from different sources. The proposed work focuses on the combination of different steps; it starts by a preprocessing step to improve the vessels appearance within the liver region of interest in the CT scans. Coherence enhancing diffusion filtering (CED) and vesselness filtering methods are used to improve vessels contrast and intensity homogeneity. The proposed U-net based network architecture is implemented with modified residual block to include concatenation skip connection. The effect of enhancement using filtering step was studied. Also, the effect of data mismatch used in training and validation is studied.
Results:
The proposed method is evaluated using many CT datasets. Dice similarity coefficient (DSC) is used to evaluate the method. The average DSC score achieved a score 79%.
Conclusions:
The proposed approach succeeded to segment liver vasculature from the liver envelope accurately, which makes it as potential tool for clinical preoperative planning.
... The deteriorated performance of the different segmentation methods of the coronary artery's vessels was due to different limitations including the calcified arteries, arteries narrowed/discontinues branches, leading to imprecise extraction of the vessels. However, the most popularly used vesselness measure includes Frangi filter, which approximates the vessel by a tubular structure [9]. ...
Diseases of coronary artery are deliberated as one of the most common heart diseases leading to death worldwide. For early detection of such disease, the X-ray angiography is a benchmark imaging modality for diagnosing such illness. The acquired X-ray angiography images usually suffer from low quality and the presence of noise. Therefore, for developing a computer-aided diagnosis (CAD) system, vessel enhancement and segmentation play significant role. In this paper, an optimized adapter filter based on Frangi filter was proposed for superior segmentation of the angiography images using genetic algorithm (GA). The original angiography image is initially preprocessed to enhance its contrast followed by generating the vesselness map using the proposed optimized Frangi filter. Then, a segmentation technique is applied to extract only the artery vessels, where the proposed system for extracting only the main vessel was evaluated. The experimental results on angiography images established the superiority of the vessel regions extraction showing 98.58% accuracy compared to the state-of-the-art.
... Hence, we set SNR as a small value (i.e., 0.8, 0.7, or 0.6) to assess the robustness of our proposed method. To be similar to the process of selection of SNR, we have studied the works (Arce-Santana et al., 2019;Jin et al., 2013) to select σ of 10, 50, and 100, respectively. ...
Prostate segmentation is an important step in prostate volume estimation, multi-modal image registration, and patient-specific anatomical modeling for surgical planning and image-guided biopsy. Manual delineation of the prostate contour is time-consuming and prone to inter- and intra-observer variability. Accurate prostate segmentation in transrectal ultrasound images is particularly challenging due to the ambiguous boundary between the prostate and neighboring organs, the presence of shadow artifacts, heterogeneous intra-prostate image intensity, and inconsistent anatomical shapes. Therefore, in this study, we propose a novel hybrid segmentation method (H-SegMed) for accurate prostate segmentation in TRUS images. The method consists of two main steps: (1) an improved closed principal curve-based method was used to obtain the data sequence, in which only few radiologist-defined seed points were used as an approximate initialization; and (2) an enhanced machine learning method was used to achieve an accurate and smooth contour of the prostate. Our results show that the proposed model achieved superior segmentation performance compared with several other state-of-the-art models, achieving an average Dice similarity coefficient, Jaccard similarity coefficient (Ω), and accuracy of 96.5, 95.1, and 96.3%, respectively.
... To solve the 3D SR problem, therefore, a model that extracts 3D structural information is preferred. In 3D models, tensors have been used to represent 3D gradient features such as in the case of vessel segmentation in medical image analysis [39]- [41]. In deep learning, 3D CNN networks [42] have achieved state-of-the-art performance with residual densely connected generators and adversarial models [43]- [45]. ...
... Tensors are generally used to obtain quantitative features such as stress tensors in mechanics and the Hessian matrix or structure tensor in image analysis [39], [41]. A tensor can be independently decomposed into shape (structure) and orientation via eigen-decomposition. ...
Surface-based analysis of magnetic resonance imaging (MRI) data of the brain plays an important role in clinical and research applications. To achieve accurate three-dimensional (3D) surface reconstruction, high-resolution (HR) MR image acquisition is needed. However, HR image acquisition is hindered by hardware limitations that result in long acquisition time and low spatial coverage. Single image super-resolution (SISR) can alleviate these problems by converting a low-resolution (LR) image to an HR image. However, unlike 2D SISR methods, conventional 3D methods incur a large computational cost and require abundant data. Further, 3D boundaries for surface reconstruction based on MR images have not been sufficiently investigated. We herein propose a cost-efficient novel regression-based framework for full 3D super-resolution imaging that directly analyzes 3D features by introducing a tensor using gradient information. We initially cluster features using tensors to create labels for both the training and testing stages. In the training stage, for each label, we collect LR patches and corresponding HR intensities to compute filters. In the testing stage, for each voxel, we construct a tensor to obtain a feature and predict the HR intensity using trained filters. We also propose a patch span reduction method by limiting patch orientation to reduce the orientation span and increase shape variety. Using only 30 masked T1-weighted brain MR volumes from the Human Connectome Project (HCP) 900 dataset, the proposed algorithm exhibited superior performance in terms of HR boundary recovery in the cerebral cortex as well as improved overall quality compared to conventional methods.
... To separate these closely packed microcolonies, we located 3D ridge-like features (e.g., blobs, sheets or tubes) that were dark relative to nearby microcolonies 76,77 . Initially, I(x, y, z) was convolved with a Gaussian kernel G(x, y, z, σ) of spatial scale σ to isolate features at that scale: H contains information on the 3D features of scale σ present within the image around position (x, y, z). ...
... Initially, I(x, y, z) was convolved with a Gaussian kernel G(x, y, z, σ) of spatial scale σ to isolate features at that scale: H contains information on the 3D features of scale σ present within the image around position (x, y, z). These features can be classified by finding the eigenvalues λ 1 , λ 2 and λ 3 of H, where λ 1 j j≥ λ 2 j j≥ λ 3 j j 76,77 . However, as we were not concerned with the specific type of feature separating microcolonies, we generated a binary segmentation of the image simply by finding all voxels for which λ 1 was above a threshold value. ...
Bacteria often live in diverse communities where the spatial arrangement of strains and species is considered critical for their ecology. However, a test of this hypothesis requires manipulation at the fine scales at which spatial structure naturally occurs. Here we develop a droplet-based printing method to arrange bacterial genotypes across a sub-millimetre array. We print strains of the gut bacterium Escherichia coli that naturally compete with one another using protein toxins. Our experiments reveal that toxin-producing strains largely eliminate susceptible non-producers when genotypes are well-mixed. However, printing strains side-by-side creates an ecological refuge where susceptible strains can persist in large numbers. Moving to competitions between toxin producers reveals that spatial structure can make the difference between one strain winning and mutual destruction. Finally, we print different potential barriers between competing strains to understand how ecological refuges form, which shows that cells closest to a toxin producer mop up the toxin and protect their clonemates. Our work provides a method to generate customised bacterial communities with defined spatial distributions, and reveals that micron-scale changes in these distributions can drive major shifts in ecology.
... Jin et al. used the hessian-based filtering with a level set active contour to segment the liver vasculature, the level set technique is modified to include the Gaussian standard deviation to reduce the effect of Gaussian filter blurring. The authors used synthetic images to evaluate their technique 10 . ...
... The constants α, β and c are constants which tune the sensitivity of, Rα, Rβ and S of the vesselness in Equation (1). In this paper the α and β are set to 0.5, the value of c depend on the gray scale range of the image, and it is proved that the optimal value of it is the half maximum value of the hessian norm 10 . ...
... The qualitative evaluation carried in the previous section justifies the segmentation method measures values. Quantitative comparisons with other segmentation techniques in the literature [6][7][8][9][10]13,26,27,33 , are hardly feasible to be held for many reasons. First, in liver vasculature segmentation subject, these techniques are few, some authors used their own clinical data sets, and some of them used synthetic data sets and others validated their work on images with constructed shapes (phantoms) similar to vessels 8 . ...
Liver vasculature segmentation is a crucial step for liver surgical planning. Segmentation of liver vasculature is an important part of the 3D visualisation of the liver anatomy. The spatial relationship between vessels and other liver structures, like tumors and liver anatomic segments, helps in reducing the surgical treatment risks. However, liver vessels segmentation is a challenging task, that is due to low contrast with neighboring parenchyma, the complex anatomy, the very thin branches and very small vessels. This paper introduces a fully automated framework consist of four steps to segment the vessels inside the liver organ. Firstly, in the preprocessing step, a combination of two filtering techniques are used to extract and enhance vessels inside the liver region, first the vesselness filter is used to extract the vessels structure, and then the anisotropic coherence enhancing diffusion (CED) filter is used to enhance the intensity within the tubular vessels structure. This step is followed by a smart multiple thresholding to extract the initial vasculature segmentation. The liver vasculature structures, including hepatic veins connected to the inferior vena cava and the portal veins, are extracted. Finally, the inferior vena cava is segmented and excluded from the vessels segmentation, as it is not considered as part of the liver vasculature structure. The liver vessel segmentation method is validated on the publically available 3DIRCAD datasets. Dice coefficient (DSC) is used to evaluate the method, the average DSC score achieved a score 68.5%. The proposed approach succeeded to segment liver vasculature from the liver envelope accurately, which makes it as potential tool for clinical preoperative planning.
... However, it was found that the Hermite shape filter and other filters analyzed in Reference [4] are unable to retain the polarities of the EMG signals. In some recent studies [20][21][22], multi-scale methods based on Hessian are proposed for enhancing line-like structure in digital images. However, none of these methods are utilized for the MUAP enhancement in the EMG images. ...
... A novel approach is presented based on the multi-scale Hessian matrix for detecting and enhancing the MUAP propagation structure in the spatio-temporal sEMG images. The proposed work is an extension of the initial work done by the authors in Reference [20] and then by the authors of References [22,23] for enhancement of linear and tubular structure in 2D and three-dimensional (3D) images. The same approach is also successfully used in Reference [21] to enhance and segment vessels in retinal digital images. ...
... The filter is based on the eigenvalues of the Hessian matrix of each pixel of sEMG spatio-temporal sEMG images. The same method is applied in References [20][21][22][23] to design a filter that enhances line-like Gaussian structures in digital images, like retinal images. Some of the methods use fixed scale and the (nonlinear) combinations of finite difference operators applied in a set of orientations for enhancement of digital images [24]. ...
Surface electromyography (sEMG) signals acquired with linear electrode array are useful in analyzing muscle anatomy and physiology. Most algorithms for signal processing, detection, and estimation require adequate quality of the input signals, however, multi-channel sEMG signals are commonly contaminated due to several noise sources. The sEMG signal needs to be enhanced prior to the digital signal and image processing to achieve the best results. This study is using spatio-temporal images to represent surface EMG signals. The motor unit action potential (MUAP) in these images looks like a linear structure, making certain angles with the x-axis, depending on the conduction velocity of the MU. A multi-scale Hessian-based filter is used to enhance the linear structure, i.e., the MUAP region, and to suppress the background noise. The proposed framework is compared with some of the existing algorithms using synthetic, simulated, and experimental sEMG signals. Results show improved detection accuracy of the motor unit action potential after the proposed enhancement as a preprocessing step.
... 22,23 The most widely used vesselness measure is based on Frangi's filter and it assumes that a vessel has an approximately tubular structure. 24 Matthias et al 27 proposed an automatic method which includes a robust threshold determination algorithm based on a histogram calculated from an automatically generated vessel tree. Here the lesions are accurately segmented and calcium scores are determined without user interaction. ...
... Hessian matrix based multiscale filtering via, Frangi's vesselness measure has been widely used for enhancement of the vessels in cardiac images. 24 Frangi's vesselness filter generates a response based on the Eigen values of Hessian matrix which in turn are obtained by the second order derivatives of the image intensities. The generated response helps in enhancing the vessel structures and suppressing the other regions. ...
... The Frangi's vesselness measure is one of the popular vessel detection approaches which uses Hessian matrix to describe the curvature of each pixel in the image. 24 Frangi's vesselness measure is derived by Equation (3). ...
Recent research suggests that the cardiovascular diseases (CVDs), seem to be the foremost cause of mortality among the world populace. Three dimensional (3D) imaging modality such as computed tomography angiography(CTA) is a standard noninvasive imaging modality which has great potentials for the visualization of heart and coronary arteries. This article presents a fully automated method for coronary artery extraction using modified Frangi's vesselness measure and region based segmentation. In this article, grayness and gradient based measures are used while computing Frangi's vesselness measure to improve the extraction of coronary arteries. The obtained vesselness measures are utilized for automatically computing the location of ostia. The locations of ostia are then used as starting seed points in region growing segmentation to extract coronary arteries. Three major coronary arteries, namely the left anterior descending artery (LAD), left circumflex artery (LCX) and right coronary artery (RCA) are segmented using the proposed method and the centerlines are extracted for the main coronary branches. The performance of the proposed method is evaluated using 12 3D CCTA data set. The experimental results reveal that during the calculation of modified Frangi's vesselness measure the proposed method gives improved results. The qualitative results obtained during the segmentation stage are also convincing. The average segmentation accuracy and overlap measure of the proposed method are 97.4% and 77.86%, respectively. Hence, the proposed automated approach can detect and extract coronary arteries in CCTA images with high performance.
... The algorithm locates the points of maximum curvature in the DEM which correspond to the surface trace of dykes. One common approach to analyze the curvature of elevation data is to use an eigenvalue analysis of the Hessian matrix [56,57]. The Hessian matrix is a 2 × 2 matrix composed of second-order partial derivatives of the input image, whereas the second-order partial derivatives are defined as a convolution with derivatives of Gaussian filter at scale σ. ...
Traditional exploration techniques usually rely on extensive field work supported by geophysical ground surveying. However, this approach can be limited by several factors such as field accessibility, financial cost, area size, climate, and public disapproval. We recommend the use of multiscale hyperspectral remote sensing to mitigate the disadvantages of traditional exploration techniques. The proposed workflow analyzes a possible target at different levels of spatial detail. This method is particularly beneficial in inaccessible and remote areas with little infrastructure, because it allows for a systematic, dense and generally noninvasive surveying. After a satellite regional reconnaissance, a target is characterized in more detail by plane-based hyperspectral mapping. Subsequently, Remotely Piloted Aircraft System (RPAS)-mounted hyperspectral sensors are deployed on selected regions of interest to provide a higher level of spatial detail. All hyperspectral data are corrected for radiometric and geometric distortions. End-member modeling and classification techniques are used for rapid and accurate lithological mapping. Validation is performed via field spectroscopy and portable XRF as well as laboratory geochemical and spectral analyses. The resulting spectral data products quickly provide relevant information on outcropping lithologies for the field teams. We show that the multiscale approach allows defining the promising areas that are further refined using RPAS-based hyperspectral imaging. We further argue that the addition of RPAS-based hyperspectral data can improve the detail of field mapping in mineral exploration, by bridging the resolution gap between airplane- and ground-based data. RPAS-based measurements can supplement and direct geological observation rapidly in the field and therefore allow better integration with in situ ground investigations. We demonstrate the efficiency of the proposed approach at the Lofdal Carbonatite Complex in Namibia, which has been previously subjected to rare earth elements exploration. The deposit is located in a remote environment and characterized by difficult terrain which limits ground surveys.
... However, general vessel segmentation cannot be used for hepatic vessel segmentation due to its high structural variations, branching complexity and small ending vessel size. For liver vasculature segmentation, a few methods have been published [81][82][83][84][85][86][87][88][89]. ...
... The level set technique is modified to include the Gaussian standard deviation to reduce the effect of Gaussian filter blurring. The authors used synthetic images to evaluate their technique; the images contained a vessel-like structure with different diameters [85]. ...
Preoperative planning for liver surgical treatments is an essential planning tool that aids in reducing the risks of surgical resection. Based on the computed tomography (CT) images, the resection can be planned before the actual tumour resection surgery. The computer-aided system provides an overview of the spatial relationships of the liver organ and its internal structures, tumours, and vasculature. It also allows for an accurate calculation of the remaining liver volume after resection. The aim of this paper was to review the main stages of the computer-aided system that helps to evaluate the risk of resection during liver cancer surgical treatments. The computer-aided system assists with surgical planning by enabling physicians to get volumetric measurements and visualise the liver, tumours, and surrounding vasculature. In this paper, it is concluded that for accurate planning of tumour resections, the liver organ and its internal structures should be segmented to understand the clear spatial relationship between them, thus allowing for a safer resection. This paper presents the main proposed segmentation techniques for each stage in the computer-aided system, namely the liver organ, tumours, and vessels. From the reviewed methods, it has been found that instead of relying on a single specific technique, a combination of a group of techniques would give more accurate segmentation results. The extracted masks from the segmentation algorithms are fused together to give the surgeons the 3D visualisation tool to study the spatial relationships of the liver and to calculate the required resection planning parameters.
