Low-contrast Images Research Articles

Deep Selective Fusion of Visible and Near-Infrared Images Using Unsupervised U-Net.

In low light conditions, visible (VIS) images are of a low dynamic range (low contrast) with severe noise and color, while near-infrared (NIR) images contain clear textures without noise and color. Multispectral fusion of VIS and NIR images produces color images of high quality, rich textures, and little noise by taking both advantages of VIS and NIR images. In this article, we propose the deep selective fusion of VIS and NIR images using unsupervised U-Net. Existing image fusion methods are afflicted with the low contrast in VIS images and flash-like effect in NIR images. Thus, we adopt unsupervised U-Net to achieve deep selective fusion of multiple scale features. Due to the absence of the ground truth, we use unsupervised learning by formulating an energy function as a loss function. To deal with insufficient training data, we perform data augmentation by rotating images and adjusting their intensity. We synthesize training data by degrading clean VIS images and masking clean NIR images using a circle. First, we utilize pretrained visual geometry group (VGG) to extract features from VIS images. Second, we build an encoding network to obtain edge information from NIR images. Finally, we combine all features and feed them into a decoding network for fusion. Experimental results demonstrate that the proposed fusion network produces visually pleasing results with fine details, little noise, and natural color and it is superior to state-of-the-art methods in terms of visual quality and quantitative measurements.

Comparing Deep Learning Performance for Chronic Lymphocytic Leukaemia Cell Segmentation in Brightfield Microscopy Images.

This article focuses on the detection of cells in low-contrast brightfield microscopy images; in our case, it is chronic lymphocytic leukaemia cells. The automatic detection of cells from brightfield time-lapse microscopic images brings new opportunities in cell morphology and migration studies; to achieve the desired results, it is advisable to use state-of-the-art image segmentation methods that not only detect the cell but also detect its boundaries with the highest possible accuracy, thus defining its shape and dimensions. We compared eight state-of-the-art neural network architectures with different backbone encoders for image data segmentation, namely U-net, U-net++, the Pyramid Attention Network, the Multi-Attention Network, LinkNet, the Feature Pyramid Network, DeepLabV3, and DeepLabV3+. The training process involved training each of these networks for 1000 epochs using the PyTorch and PyTorch Lightning libraries. For instance segmentation, the watershed algorithm and three-class image semantic segmentation were used. We also used StarDist, a deep learning-based tool for object detection with star-convex shapes. The optimal combination for semantic segmentation was the U-net++ architecture with a ResNeSt-269 background with a data set intersection over a union score of 0.8902. For the cell characteristics examined (area, circularity, solidity, perimeter, radius, and shape index), the difference in mean value using different chronic lymphocytic leukaemia cell segmentation approaches appeared to be statistically significant (Mann-Whitney U test, P < .0001). We found that overall, the algorithms demonstrate equal agreement with ground truth, but with the comparison, it can be seen that the different approaches prefer different morphological features of the cells. Consequently, choosing the most suitable method for instance-based cell segmentation depends on the particular application, namely, the specific cellular traits being investigated.

Accurate Airway Tree Segmentation in CT Scans via Anatomy-aware Multi-class Segmentation and Topology-guided Iterative Learning.

Intrathoracic airway segmentation in computed tomography is a prerequisite for various respiratory disease analyses such as chronic obstructive pulmonary disease, asthma and lung cancer. Due to the low imaging contrast and noises execrated at peripheral branches, the topological-complexity and the intra-class imbalance of airway tree, it remains challenging for deep learning-based methods to segment the complete airway tree (on extracting deeper branches). Unlike other organs with simpler shapes or topology, the airway's complex tree structure imposes an unbearable burden to generate the "ground truth" label (up to 7 or 3 hours of manual or semi-automatic annotation per case). Most of the existing airway datasets are incompletely labeled/annotated, thus limiting the completeness of computer-segmented airway. In this paper, we propose a new anatomy-aware multi-class airway segmentation method enhanced by topology-guided iterative self-learning. Based on the natural airway anatomy, we formulate a simple yet highly effective anatomy-aware multi-class segmentation task to intuitively handle the severe intra-class imbalance of the airway. To solve the incomplete labeling issue, we propose a tailored iterative self-learning scheme to segment toward the complete airway tree. For generating pseudo-labels to achieve higher sensitivity (while retaining similar specificity), we introduce a novel breakage attention map and design a topology-guided pseudo-label refinement method by iteratively connecting breaking branches commonly existed from initial pseudo-labels. Extensive experiments have been conducted on four datasets including two public challenges. The proposed method achieves the top performance in both EXACT'09 challenge using average score and ATM'22 challenge on weighted average score. In a public BAS dataset and a private lung cancer dataset, our method significantly improves previous leading approaches by extracting at least (absolute) 6.1% more detected tree length and 5.2% more tree branches, while maintaining comparable precision.

Rethinking automatic segmentation of gross target volume from a decoupling perspective

Accurate and reliable segmentation of Gross Target Volume (GTV) is critical in cancer Radiation Therapy (RT) planning, but manual delineation is time-consuming and subject to inter-observer variations. Recently, deep learning methods have achieved remarkable success in medical image segmentation. However, due to the low image contrast and extreme pixel imbalance between GTV and adjacent tissues, most existing methods usually obtained limited performance on automatic GTV segmentation. In this paper, we propose a Heterogeneous Cascade Framework (HCF) from a decoupling perspective, which decomposes the GTV segmentation into independent recognition and segmentation subtasks. The former aims to screen out the abnormal slices containing GTV, while the latter performs pixel-wise segmentation of these slices. With the decoupled two-stage framework, we can efficiently filter normal slices to reduce false positives. To further improve the segmentation performance, we design a multi-level Spatial Alignment Network (SANet) based on the feature pyramid structure, which introduces a spatial alignment module into the decoder to compensate for the information loss caused by downsampling. Moreover, we propose a Combined Regularization (CR) loss and Balance-Sampling Strategy (BSS) to alleviate the pixel imbalance problem and improve network convergence. Extensive experiments on two public datasets of StructSeg2019 challenge demonstrate that our method outperforms state-of-the-art methods, especially with significant advantages in reducing false positives and accurately segmenting small objects. The code is available at https://github.com/shijun18/GTV_AutoSeg.

Spreading depolarization suppression from inter-astrocytic gap junction blockade assessed with multimodal imaging and a novel wavefront detection scheme

Spreading depolarizations (SDs) are an enigmatic and ubiquitous co-morbidity of neural dysfunction. SDs are propagating waves of local field depolarization and increased extracellular potassium. They increase the metabolic demand on brain tissue, resulting in changes in tissue blood flow, and are associated with adverse neurological consequences including stroke, epilepsy, neurotrauma, and migraine. Their occurrence is associated with poor patient prognosis through mechanisms which are only partially understood. Here we show in vivo that two (structurally dissimilar) drugs, which suppress astroglial gap junctional communication, can acutely suppress SDs. We found that mefloquine hydrochloride (MQH), administered IP, slowed the propagation of the SD potassium waveform and intermittently led to its suppression. The hemodynamic response was similarly delayed and intermittently suppressed. Furthermore, in instances where SD led to transient tissue swelling, MQH reduced observable tissue displacement. Administration of meclofenamic acid (MFA) IP was found to reduce blood flow, both proximal and distal, to the site of SD induction, preceding a large reduction in the amplitude of the SD-associated potassium wave. We introduce a novel image processing scheme for SD wavefront localization under low-contrast imaging conditions permitting full-field wavefront velocity mapping and wavefront parametrization. We found that MQH administration delayed SD wavefront's optical correlates. These two clinically used drugs, both gap junctional blockers found to distinctly suppress SDs, may be of therapeutic benefit in the various brain disorders associated with recurrent SDs.

Underwater Image Enhancement Based on Color Feature Fusion

The ever-changing underwater environment, coupled with the complex degradation modes of underwater images, poses numerous challenges to underwater image enhancement efforts. Addressing the issues of low contrast and significant color deviations in underwater images, this paper presents an underwater image enhancement approach based on color feature fusion. By leveraging the properties of light propagation underwater, the proposed model employs a multi-channel feature extraction strategy, using convolution blocks of varying sizes to extract features from the red, green, and blue channels, thus effectively learning both global and local information of underwater images. Moreover, an attention mechanism is incorporated to design a residual enhancement module, augmenting the capability of feature representation. Lastly, a dynamic feature enhancement module is designed using deformable convolutions, enabling the network to capture underwater scene information with higher precision. Experimental results on public datasets demonstrate the outstanding performance of our proposed method in underwater image enhancement. Further, object detection experiments conducted on pre- and post-enhanced images underscore the value of our method for downstream tasks.

Thermal-visible stereo matching at night based on Multi-Modal Autoencoder

Compared to RGB cameras, thermal sensors are more reliable at night. However, in the stereo matching task, it is difficult to estimate reliable depth simply depending on the binocular thermal cameras, due to their monochrome and low-contrast images. This paper proposed a thermal-visible stereo matching method based on multi-modal autoencoder (MANet), the key to which is to extract the modal-invariant features between thermal and visible images, and meanwhile, retain the characteristic features in terms of the optical properties. Besides, we designed a cross-reconstruction constraint to improve the scene feature extraction capability of MANet. In the stereo matching process, we analyze the thermal and visible properties and adopt optical attention mechanism to improve night-scene stereo matching. We established a stereo matching system with thermal-visible camera based on LiDAR calibration, and provided the dataset for evaluation. The experiment demonstrates that the proposed MANet can effectively extract the modal-invariant features and can be applied on thermal-visible stereo matching.

RMAP-ResNet: Segmentation of brain tumor OCT images using residual multicore attention pooling networks for intelligent minimally invasive theranostics

Background and objectiveThe occurrence of brain tumors seriously affects patients both mentally and physically, reducing their quality of life. Automatic recognition and segmentation of tumor boundaries are essential for computer-aided diagnosis and treatment. While optical coherence tomography (OCT) has been widely employed in the detection and analysis fundus lesions and vascular segmentation due to its high resolution and real-time imaging advantages, the segmentation of brain tumor OCT images based on deep neural network has not been studied. However, despite the encoding and decoding structure based on deep learning has achieved good results in medical image segmentation, these models still face limitations in accurately segmenting tumor OCT images due to variations in tumor shapes, sizes, and numbers, and the low contrast of OCT images. Accurate segmentation of tumor tissue contours is crucial as the absence of this mechanism hampers the effective application of OCT imaging methods for brain tumor diagnosis and treatment in clinical practice. Therefore, we propose a novel neural network that can accurately segment brain tumors and their residual tissues in OCT images. MethodsIn this study, we propose a novel residual multi-core attention pooling network (RMAP ResNet) for the segmentation of tumorous and its residual tissue. The RMAP module primarily relies on multiple associative effective receptive fields. These fields enable the detection of targets of various sizes. The module also incorporates an attention structure to adjust connection weights, suppressing irrelevant information and emphasizing the characteristics of tumor OCT images. This process restores the contextual dependencies. ResultsWe validated our proposed RMAP ResNet method on the dataset of mouse tumor OCT imaging obtained using the sweep source OCT imaging system. The experimental results demonstrate that RMAP ResNet achieves superior performance in tumor tissue segmentation. The proposed method improves the quality of tumor segmentation for various shapes and sizes, leading to superior results. RMAP ResNet achieved Dice and IoU values of 94.78% and 90.58%, respectively, confirming the effectiveness and advancement of our proposed technology. ConclusionsOur proposed method addresses the crucial issue of the limited use of the OCT imaging modality in tumor resection surgery. It overcomes the challenge of visually identifying the tumor boundary in OCT modality imaging, where physicians using the naked eye may encounter difficulties. This study has potential applications in the fields of surgical navigation and intelligent minimally invasive treatment of brain tumors.

Adaptive high-gray image enhancement algorithm based on logarithmic mapping and simulated exposure

In the fields of industry, security, military and other areas, high gray-level low contrast images have been widely used. However, due to the large dynamic range of gray level, high noise, blurry detail information and low contrast, it brings difficulties to subsequent image enhancement, target detection and other processes. Ordinary image enhancement algorithms are only suitable for conventional 8-bit images, which are not effective to deal with high-gray and low-contrast images. To address these issues, this paper proposed an adaptive high-gray image enhancement algorithm based on logarithmic mapping and simulated exposure, which was used for processing X-ray film images, infrared images, SAR images and other high gray-level low contrast images. Firstly, a detailed algorithm model construction method was provided for high gray-level low contrast images. Secondly, in order to verify the universality of the algorithm proposed in this paper, qualitative and quantitative experimental analyses were carried out using X-ray film images, infrared images and SAR images respectively. Finally, taking the most complex infrared image in the application environment as an example, the images, enhanced by the algorithm proposed in this paper, were sent to the YOLOv8 neural network for target detection, which improved the detection accuracy. The experiment shows that the algorithm, proposed in this paper, can significantly improve the image quality of high gray-level low contrast images, the detail information is obvious and the contrast is significantly improved. Various indicators are far higher than those of traditional algorithms. At the same time, this algorithm can provide reliable technical support for applications, such as target detection and target tracking in this field.

TANet: Transmission and atmospheric light driven enhancement of underwater images

Motivated by the adverse impact of light attenuation and scattering, which leads to color distortion and low contrast in underwater images, our study primarily focuses on enhancement techniques for these images using localized transmission feature analysis and global atmospheric light feature extraction. To this end, we propose a novel approach, named TANet, drawing upon the dynamics of transmission and atmospheric light. TANet integrates two primary components: a spatial domain-based Transmission-Driven Refinement module (TDR) and a frequency domain-based Atmospheric Light Removal Fourier Module (ALRF). The TDR module employs a Gated Multipurpose Unit with dual branches, selectively regulating input features. This allows for a refined merging of feature vectors that subsequently interact, enabling cross-channel feature integration. Capitalizing on the correlation between transmission and image quality, TDR facilitates the detailed enhancement of underwater images by depicting the perceived transmission across distinct image sections. Given that atmospheric light exhibits different attenuation rates under water due to varying wavelengths, and considering that atmospheric light is globally constant, thereby influencing underwater image capture, we developed the ALRF module. This caters to the processing of global information within the frequency domain, efficiently negating atmospheric light’s impact on underwater images and enhancing their quality and visibility. Our TANet’s superior performance is affirmed by extensive experimental results, demonstrating its effectiveness in underwater image enhancement.

A fractional-order image segmentation model with application to low-contrast and piecewise smooth images

In this paper, we propose a two-stage image segmentation model based on structure tensor and fractional-order regularization. In the first stage, we use the fractional-order regularization to approximate the Hausdorff measure of the Mumford-Shah (MS) model. The existence and uniqueness of the solution is proved and the alternating direction implicit (ADI) scheme is used to find the solution of the modified MS model. In the second stage, a thresholding is used to induce the segmentation of the target. The superior performances of the proposed model are demonstrated by some comparative experimental results with several state-of-art methods.

Deep learning image segmentation for the reliable porosity measurement of high-capacity Ni-based oxide cathode secondary particles

The optimization of geometrical pore control in high-capacity Ni-based cathode materials is required to enhance the cyclic performance of lithium-ion batteries. Enhanced porosity improves lithium-ion mobility by increasing the electrode–electrolyte contact area and reducing the number of ion diffusion pathways. However, excessive porosity can diminish capacity, thus necessitating optimizing pore distribution to compromise the trade-off relation. Accordingly, a statistically meaningful porosity estimation of electrode materials is required to engineer the local pore distribution inside the electrode particles. Conventional scanning electron microscopy (SEM) image-based porosity measurement can be used for this purpose. However, it is labor-intensive and subjected to human bias for low-contrast pore images, thereby potentially lowering measurement accuracy. To mitigate these difficulties, we propose an automated image segmentation method for the reliable porosity measurement of cathode materials using deep convolutional neural networks specifically trained for the analysis of porous cathode materials. Combined with the preprocessed SEM image datasets, the model trained for 100 epochs exhibits an accuracy of > 97% for feature segmentation with regard to pore detection on the input datasets. This automated method considerably reduces manual effort and human bias related to the digitization of pore features in serial section SEM image datasets used in 3D electron tomography.Graphical abstract

Multimodal image feature fusion for improving medical ultrasound image segmentation

Ultrasound imaging is the most commonly used technique to detect abnormalities of the thyroid and diagnose abnormal breasts. However, due to the inherent characteristics of ultrasound imaging, the contrast of ultrasound images is relatively low, and the tissue boundaries in the image are blurry. This makes it a challenging task to segment thyroid or breast lesions from ultrasound images. In this work, we explore the role of multimodal features at different levels of deep neural network models and their fusion mechanisms for multimodal images generated by ultrasound radio frequency (RF) data. The ultrasound RF data contain information related to the arrangement of scatterers in tissues and organs, which is only partially available in B-mode images. To solve this problem, we use ultrasound RF data to generate parametric images, including Nakagami and entropy images, to illustrate the abundant backscattered statistics contained in RF data. A channel-aware fusion module is proposed to adaptively fuse the features from images of different modalities. Experimental results show that the abundant information contained in RF data can be helpful to identify thyroid glands or breast lesions with low contrast and fuzzy boundaries in B-mode ultrasound images and improve the segmentation performance.

Contrast Adaptation in Pseudophakic Patients with Macular Disorders

Purpose To learn whether contrast adaptation, induced by positive spherical defocus, is compromised by macular disorders such as age-related macular degeneration (AMD) or epiretinal membranes (ERM) and to gain further insight in the functionality of the pathological macula and the level of “functional reserve” often postulated for the indication of presbyopia correcting IOLs. Methods In a pilot study, patients with macular disorders, AMD and ERM, (n = 10) and healthy volunteers (n = 10) were tested to quantify contrast adaption after +4 D defocus for 10 min, by performing an interocular contrast matching task. Subjects manually adjusted the perceived contrast of the test patch as seen by the test eye to match to the contrast of a target patch with a fixed Michelson contrast of 0.2 as seen by the contralateral untreated eye. Results Patients with macular disorders subjectively matched the 0.2 target contrast with a contrast of 0.24 ± 0.06 (mean ± SD) before adaptation and with a contrast of 0.19 ± 0.04 after adaptation (p < 0.05). Accordingly, patients with macular disorders showed an induced contrast gain by adaptation of 0.05 (27%), which was not different from the control group, which showed an induced contrast gain by adaptation of 0.06 (35%). Conclusion Patients with mild and moderate macular disorders, such as AMD and ERM, show an induced contrast adaptation, i.e. a gain in contrast sensitivity, at 3.2 cpd, which is not different in level from the induced contrast adaptation in healthy subjects. Macular disorders do not prevent adaptation of the patient’s visual system to low contrast or blurred retinal images. Therefore, the implantation of presbyopia correcting IOLs is not a strict exclusion criterion for these patients, but the progressive nature of the macular disorder must be taken into account.

Brain tumor feature extraction and edge enhancement algorithm based on U-Net network

BackgroundStatistics show that each year more than 100,000 patients pass away from brain tumors. Due to the diverse morphology, hazy boundaries, or unbalanced categories of medical data lesions, segmentation prediction of brain tumors has significant challenges. PurposeIn this thesis, we highlight EAV-UNet, a system designed to accurately detect lesion regions. Optimizing feature extraction, utilizing automatic segmentation techniques to detect anomalous regions, and strengthening the structure. We prioritize the segmentation problem of lesion regions, especially in cases where the margins of the tumor are more hazy. MethodsThe VGG-19 network structure is incorporated into the coding stage of the U-Net, resulting in a deeper network structure, and an attention mechanism module is introduced to augment the feature information. Additionally, an edge detection module is added to the encoder to extract edge information in the image, which is then passed to the decoder to aid in reconstructing the original image. Our method uses the VGG-19 in place of the U-Net encoder. To strengthen feature details, we integrate a CBAM (Channel and Spatial Attention Mechanism) module into the decoder to enhance it. To extract vital edge details from the data, we incorporate an edge recognition section into the encoder. ResultsAll evaluation metrics show major improvements with our recommended EAV-UNet technique, which is based on a thorough analysis of experimental data. Specifically, for low contrast and blurry lesion edge images, the EAV-Unet method consistently produces forecasts that are very similar to the initial images. This technique reduced the Hausdorff distance to 1.82, achieved an F1 score of 96.1%, and attained a precision of 93.2% on Dataset 1. It obtained an F1 score of 76.8%, a Precision of 85.3%, and a Hausdorff distance reduction to 1.31 on Dataset 2. Dataset 3 displayed a Hausdorff distance cut in 2.30, an F1 score of 86.9%, and Precision of 95.3%. ConclusionsWe conducted extensive segmentation experiments using various datasets related to brain tumors. We refined the network architecture by employing smaller convolutional kernels in our strategy. To further improve segmentation accuracy, we integrated attention modules and an edge enhancement module to reinforce edge information and boost attention scores.

Unsupervised deep network for image texture transformation: Improving the quality of cross-correlation analysis and mechanical vortex visualisation during cardiac fibrillation

Visualisation of cardiac fibrillation plays a very considerable role in cardiophysiological study and clinical applications. One of the ways to obtain the image of these phenomena is the registration of mechanical displacement fields reflecting the track from electrical activity. In this work, we read these fields using cross-correlation analysis from the video of open pig's epicardium at the start of fibrillation recorded with electrocardiogram. However, the quality of obtained displacement fields remains low due to the weak pixels heterogeneity of the frames. It disables to see more clearly such interesting phenomena as mechanical vortexes that underline the mechanical dysfunction of fibrillation. The applying of chemical or mechanical markers to solve this problem can affect the course of natural processes and falsify the results. Therefore, we developed a novel scheme of an unsupervised deep neural network that is based on the state-of-art positional coding technology for a multilayer perceptron. This network enables to generate a couple of frames with a more heterogeneous pixel texture, that is more suitable for cross-correlation analysis methods, from two consecutive frames. The novel network scheme was tested on synthetic pairs of images with different texture heterogeneity and frequency of displacement fields and also it was compared with different filters on our cardiac tissue image dataset. The testing showed that the displacement fields obtained with our method are closer to the ground truth than those which were computed only with the cross-correlation analysis in low contrast images case where filtering is impossible. Moreover, our model showed the best results comparing with the one of the popular filter CLAHE on our dataset. As a result, using our approach, it was possible to register more clearly a mechanical vortex on the epicardium at the start of fibrillation continuously for several milliseconds for the first time.

Comparison of image quality and quantitative parameters in intravoxel incoherent motion imaging at 3-T based on turbo spin-echo and echo-planar imaging in patients with oral cancer.

To compare the image quality, apparent diffusion coefficient (ADC), and intravoxel incoherent motion- (IVIM) derived parameters of IVIM imaging based on turbo spin-echo (TSE) and echo-planar imaging (EPI) of patients with oral cancer and to assess the equivalence of the ADC and IVIM-derived parameters. Thirty patients with oral cancer underwent TSE-IVIM and EPI-IVIM imaging using a 3.0-T system. The distortion ratio (DR), signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), qualitative evaluations of image quality, ADC, pure diffusion coefficient (D), pseudo-diffusion coefficient (D*), and perfusion fraction (f) were compared between the two sequences. The consistency of the quantitative parameters in oral cancer between the TSE and EPI sequences was evaluated using a Bland-Altman analysis. TSE-IVIM had a significantly smaller DR than EPI-IVIM (P < 0.001). The CNR of EPI-IVIM on most of the anatomical sites was significantly higher than that of TSE-IVIM (P < 0.05), while the SNR was not significantly different (P > 0.05). TSE-IVIM had significantly higher image quality, less distortion and artifacts, and lower image contrast compared with EPI-IVIM (P < 0.05). The lesion-edge sharpness and diagnostic confidence of EPI-IVIM were lower than that of TSE-IVIM, although no significant differences existed (P > 0.05). The ADC and D of TSE-IVIM had better reproducibility (intraclass correlation coefficient > 0.9). Although no significant difference existed for the ADC and IVIM-derived parameters of lesions between the two sequences (P > 0.05), wide limits of agreement were found in the Bland-Altman plots. TSE-IVIM could be used as an alternative technique to EPI-IVIM for patients with oral cancer because of its better image quality. Furthermore, TSE-IVIM can provide more accurate quantitative parameters. However, the quantitative parameters derived from the two IVIM techniques cannot be used as equivalent parameters for patients with oral cancer.

Label-free grain segmentation for optical microscopy images via unsupervised image-to-image translation

Grain boundaries play an important role in governing the mechanical and physical properties of polycrystalline materials. Therefore, quantitative analysis of grain structure is an important prerequisite for establishing structure-property relationships. However, this quantitative analysis is hampered by the low contrast and non-uniform illumination of optical microscopy images. Although previous studies have used neural networks to detect grain boundaries in optical micrographs, their practical applications have been restricted due to labeling costs for supervised learning and their sensitivity to defects present in a grain structure image. In this paper, we approach grain boundary detection as a real-to-virtual (R2V) translation problem, mapping a single real microstructure to virtual microstructures. With this perspective, our framework benefits from two learning schemes: unsupervised learning and physics-inspired learning. Extensive experiments demonstrate the superiority and generality of our framework. We expect that our approach will facilitate the use of data-driven techniques in the field of quantitative microstructure analysis.

A multispectral transmission image cluster analysis method based on “Terrace compression Method” and window function

Multispectral transmission imaging has a great potential for the early screening of breast cancer due to the low cost, safety and ease in operation. Accurate detection of heterogeneity is important for the diagnosis of breast disease. The low contrast and unclear heterogeneity boundaries of transmission images can lead to difficulties in recognition and segmentation. Therefore, we propose a clustering segmentation method of multispectral transmission images based on “terrace compression method” and window transformation. The images are preprocessed by the frame accumulation to improve the signal-to-noise ratio. After that, the “terrace compression method” is used to compress the images nonlinearly, to reduce data redundancy and then to improve the edge information of heterogeneities. Afterwards, the window function is used to eliminate the redundant information of the background and to reduce the influence of background noise on clustering. Finally, the processed images at each wavelength are transformed into multidimensional data for cluster analysis. The multispectral transmission images of breast phantom are acquired for experimental validation. Then, compared the method with common clustering segmentation methods (including K-means, K-means++, Mean-shift, Gaussian Mixture). The results showed that this processing method can effectively segment and classify the three heterogeneities in the breast phantom. Among these methods, the method proposed in the paper has the best segmentation and classification results for the three types of heterogeneities in the breast phantom. The Dice coefficients of all the heterogeneities segmentation reached more than 0.84 and increased by a maximum of 1.08 times as compared to the common clustering methods. The applications of the terrace compression method and the grayscale window transformation improved the effect of image clustering segmentation.

تجربة عن الاستشهادات

We demonstrate a method to extract the heart outlines from noisy echocardiograph data in this paper. In order to eliminate significant noise and obtain improved image quality, our method is based on enhancing the image before applying contour detection. To accomplish that, we combine a variety of pre-processing techniques (filtering, morphological operations, and contrast adjustment) to enhance low contrast in echocardiograph images and avoid unclear edges. After applying these techniques, we are able to detect the boundaries of the heart and the movement of the valves using conventional edge detection techniques.