Deep Learning-based Segmentation Research Articles

Deep Learning-Based Segmentation and Classification Techniques for Brain Tumor MRI: A Review

Early detection of brain tumors is critical for enhancing treatment options and extending patient survival. Magnetic resonance imaging (MRI) scanning gives more detailed information, such as greater contrast and clarity than any other scanning method. Manually dividing brain tumors from many MRI images collected in clinical practice for cancer diagnosis is a tough and time-consuming task. Tumors and MRI scans of the brain can be discovered using algorithms and machine learning technologies, making the process easier for doctors because MRI images can appear healthy when the person may have a tumor or be malignant. Recently, deep learning techniques based on deep convolutional neural networks have been used to analyze medical images with favorable results. It can help save lives faster and rectify some medical errors. In this study, we look at the most up-to-date methodologies for medical image analytics that use convolutional neural networks on MRI images. There are several approaches to diagnosing and classifying brain cancers. Inside the brain, irregular cells grow so that a brain tumor appears. The size of the tumor and the part of the brain affected impact the symptoms.

Unsupervised Representation Learning for Tissue Segmentation in Histopathological Images: From Global to Local Contrast.

Tissue segmentation is an essential task in computational pathology. However, relevant datasets for such a pixel-level classification task are hard to obtain due to the difficulty of annotation, bringing obstacles for training a deep learning-based segmentation model. Recently, contrastive learning has provided a feasible solution for mitigating the heavy reliance of deep learning models on annotation. Nevertheless, applying contrastive loss to the most abstract image representations, existing contrastive learning frameworks focus on global features, therefore, are less capable of encoding finer-grained features (e.g., pixel-level discrimination) for the tissue segmentation task. Enlightened by domain knowledge, we design three contrastive learning tasks with multi-granularity views (from global to local) for encoding necessary features into representations without accessing annotations. Specifically, we construct: (1) an image-level task to capture the difference between tissue components, i.e., encoding the component discrimination; (2) a superpixel-level task to learn discriminative representations of local regions with different tissue components, i.e., encoding the prototype discrimination; (3) a pixel-level task to encourage similar representations of different tissue components within a local region, i.e., encoding the spatial smoothness. Through our global-to-local pre-training strategy, the learned representations can reasonably capture the domain-specific and fine-grained patterns, making them easily transferable to various tissue segmentation tasks in histopathological images. We conduct extensive experiments on two tissue segmentation datasets, while considering two real-world scenarios with limited or sparse annotations. The experimental results demonstrate that our framework is superior to existing contrastive learning methods and can be easily combined with weakly supervised and semi-supervised segmentation methods.

Segmentation of biventricle in cardiac cine MRI via nested capsule dense network.

Cardiac magnetic resonance image (MRI) has been widely used in diagnosis of cardiovascular diseases because of its noninvasive nature and high image quality. The evaluation standard of physiological indexes in cardiac diagnosis is essentially the accuracy of segmentation of left ventricle (LV) and right ventricle (RV) in cardiac MRI. The traditional symmetric single codec network structure such as U-Net tends to expand the number of channels to make up for lost information that results in the network looking cumbersome. Instead of a single codec, we propose a multiple codecs structure based on the FC-DenseNet (FCD) model and capsule convolution-capsule deconvolution, named Nested Capsule Dense Network (NCDN). NCDN uses multiple codecs to achieve multi-resolution, which makes it possible to save more spatial information and improve the robustness of the model. The proposed model is tested on three datasets that include the York University Cardiac MRI dataset, Automated Cardiac Diagnosis Challenge (ACDC-2017), and the local dataset. The results show that the proposed NCDN outperforms most methods. In particular, we achieved nearly the most advanced accuracy performance in the ACDC-2017 segmentation challenge. This means that our method is a reliable segmentation method, which is conducive to the application of deep learning-based segmentation methods in the field of medical image segmentation.

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography.

Recently, deep learning-based segmentation models have been widely applied in the ophthalmic field. This study presents the complete process of constructing an orbital computed tomography (CT) segmentation model based on U-Net. For supervised learning, a labor-intensive and time-consuming process is required. The method of labeling with super-resolution to efficiently mask the ground truth on orbital CT images is introduced. Also, the volume of interest is cropped as part of the pre-processing of the dataset. Then, after extracting the volumes of interest of the orbital structures, the model for segmenting the key structures of the orbital CT is constructed using U-Net, with sequential 2D slices that are used as inputs and two bi-directional convolutional long-term short memories for conserving the inter-slice correlations. This study primarily focuses on the segmentation of the eyeball, optic nerve, and extraocular muscles. The evaluation of the segmentation reveals the potential application of segmentation to orbital CT images using deep learning methods.

MicrobeSEG: A deep learning software tool with OMERO data management for efficient and accurate cell segmentation.

In biotechnology, cell growth is one of the most important properties for the characterization and optimization of microbial cultures. Novel live-cell imaging methods are leading to an ever better understanding of cell cultures and their development. The key to analyzing acquired data is accurate and automated cell segmentation at the single-cell level. Therefore, we present microbeSEG, a user-friendly Python-based cell segmentation tool with a graphical user interface and OMERO data management. microbeSEG utilizes a state-of-the-art deep learning-based segmentation method and can be used for instance segmentation of a wide range of cell morphologies and imaging techniques, e.g., phase contrast or fluorescence microscopy. The main focus of microbeSEG is a comprehensible, easy, efficient, and complete workflow from the creation of training data to the final application of the trained segmentation model. We demonstrate that accurate cell segmentation results can be obtained within 45 minutes of user time. Utilizing public segmentation datasets or pre-labeling further accelerates the microbeSEG workflow. This opens the door for accurate and efficient data analysis of microbial cultures.

Deep learning-based image analysis reveals significant differences in the number and distribution of mucosal CD3 and γδ T cells between Crohn's disease and ulcerative colitis.

Colon mucosae of ulcerative colitis (UC) and Crohn's disease (CD) display differences in the number and distribution of immune cells that are difficult to assess by eye. Deep learning-based analysis on whole slide images (WSIs) allows extraction of complex quantitative data that can be used to uncover different inflammatory patterns. We aimed to explore the distribution of CD3 and γδ T cells in colon mucosal compartments in histologically inactive and active inflammatory bowel disease. By deep learning-based segmentation and cell detection on WSIs from a well-defined cohort of CD (n=37), UC (n=58), and healthy controls (HCs, n=33), we quantified CD3 and γδ T cells within and beneath the epithelium and in lamina propria in proximal and distal colon mucosa, defined by the Nancy histological index. We found that inactive CD had significantly fewer intraepithelial γδ T cells than inactive UC, but higher total number of CD3 cells in all compartments than UC and HCs. Disease activity was associated with a massive loss of intraepithelial γδ T cells in UC, but not in CD. The total intraepithelial number of CD3 cells remained constant regardless of disease activity in both CD and UC. There were more mucosal CD3 and γδ T cells in proximal versus distal colon. Oral corticosteroids had an impact on γδ T cell numbers, while age, gender, and disease duration did not. Relative abundance of γδ T cells in mucosa and blood did not correlate. This study reveals significant differences in the total number of CD3 and γδ T cells in particularly the epithelial area between CD, UC, and HCs, and demonstrates useful application of deep segmentation to quantify cells in mucosal compartments.

Sarcopenia identified by computed tomography imaging using a deep learning-based segmentation approach impacts survival in patients with newly diagnosed multiple myeloma.

Sarcopenia increases with age and is associated with poor survival outcomes in patients with cancer. By using a deep learning-based segmentation approach, clinical computed tomography (CT) images of the abdomen of patients with newly diagnosed multiple myeloma (NDMM) were reviewed to determine whether the presence of sarcopenia had any prognostic value. Sarcopenia was detected by accurate segmentation and measurement of the skeletal muscle components present at the level of the L3 vertebrae. These skeletal muscle measurements were further normalized by the height of the patient to obtain the skeletal muscle index for each patient to classify them as sarcopenic or not. The study cohort consisted of 322 patients of which 67 (28%) were categorized as having high risk (HR) fluorescence in situ hybridization (FISH) cytogenetics. A total of 171 (53%) patients were sarcopenic based on their peri-diagnosis standard-dose CT scan. The median overall survival (OS) and 2-year mortality rate for sarcopenic patients was 44months and 40% compared to 90months and 18% for those not sarcopenic, respectively (p<.0001 for both comparisons). In a multivariable model, the adverse prognostic impact of sarcopenia was independent of International Staging System stage, age, and HR FISH cytogenetics. Sarcopenia identified by a machine learning-based convolutional neural network algorithm significantly affects OS in patients with NDMM. Future studies using this machine learning-based methodology of assessing sarcopenia in larger prospective clinical trials are required to validate these findings.

Deep Learning-Based Segmentation of Head and Neck Organs-at-Risk with Clinical Partially Labeled Data.

Radiotherapy is one of the main treatments for localized head and neck (HN) cancer. To design a personalized treatment with reduced radio-induced toxicity, accurate delineation of organs at risk (OAR) is a crucial step. Manual delineation is time- and labor-consuming, as well as observer-dependent. Deep learning (DL) based segmentation has proven to overcome some of these limitations, but requires large databases of homogeneously contoured image sets for robust training. However, these are not easily obtained from the standard clinical protocols as the OARs delineated may vary depending on the patient's tumor site and specific treatment plan. This results in incomplete or partially labeled data. This paper presents a solution to train a robust DL-based automated segmentation tool exploiting a clinical partially labeled dataset. We propose a two-step workflow for OAR segmentation: first, we developed longitudinal OAR-specific 3D segmentation models for pseudo-contour generation, completing the missing contours for some patients; with all OAR available, we trained a multi-class 3D convolutional neural network (nnU-Net) for final OAR segmentation. Results obtained in 44 independent datasets showed superior performance of the proposed methodology for the segmentation of fifteen OARs, with an average Dice score coefficient and surface Dice similarity coefficient of 80.59% and 88.74%. We demonstrated that the model can be straightforwardly integrated into the clinical workflow for standard and adaptive radiotherapy.

E-MPSPNet: Ice–Water SAR Scene Segmentation Based on Multi-Scale Semantic Features and Edge Supervision

Distinguishing sea ice and water is crucial for safe navigation and carrying out offshore activities in ice zones. However, due to the complexity and dynamics of the ice–water boundary, it is difficult for many deep learning-based segmentation algorithms to achieve accurate ice–water segmentation in synthetic aperture radar (SAR) images. In this paper, we propose an ice–water SAR segmentation network, E-MPSPNet, which can provide effective ice–water segmentation by fusing semantic features and edge information. The E-MPSPNet introduces a multi-scale attention mechanism to better fuse the ice–water semantic features and designs an edge supervision module (ESM) to learn ice–water edge features. The ESM not only provides ice–water edge prediction but also imposes constraints on the semantic feature extraction to better express the edge information. We also design a loss function that focuses on both ice–water edges and semantic segmentations of ice and water for overall network optimization. With the AI4Arctic/ASIP Sea Ice Dataset as the benchmark, experimental results show our E-MPSPNet achieves the best performance compared with other commonly used segmentation models, reaching 94.2% for accuracy, 93.0% for F-score, and 89.2% for MIoU. Moreover, our E-MPSPNet shows a relatively smaller model size and faster processing speed. The application of the E-MPSPNet for processing a SAR scene demonstrates its potential for operational use in drawing near real-time navigation charts of sea ice.

Abstract 15482: Self-Adaptive Deep Learning-Based Segmentation for Uniform and Functional Preclinical and Clinical CT Image Analysis

Currently, effective treatment options for heart failure are lacking. Methods to monitor cardiac function longitudinally can accelerate preclinical and clinical research into disease progression and novel drugs. However, manual image analysis of cardiac substructures is resource-intensive and error-prone. While automated methods exist for clinical CT images, translating these to preclinical μCT data is challenging. We aimed at employing deep-learning to automatically extract quantitative data on cardiac functional parameters, developing a uniform method for both μCT and CT images. For this, we collected 110 contrast-enhanced μCT images of wild-type and accelerated aging (Ercc1 Δ/- ) mice, as well as a public dataset of 60 segmented cardiac CT images. Substructures, i.e., the left ventricle, left myocardium and right ventricle, were manually segmented in 24 μCT images to create a training set. After template-based heart detection, two separate segmentation neural network architectures were trained using the nnU-Net framework. Automatic and manual segmentations of the μCT training set were nearly identical (mean Dice score 0.983). The estimated median Dice score (0.940) of the test set results was comparable to existing methods for μCT image segmentation. Also, the automatic volume metrics were similar to manual observations by an expert. In Ercc1 Δ/- mice, ejection fractions had significantly decreased, and myocardial mass increased by age 24 weeks. The mean Dice score of the CT segmentation results (0.925 ± 0.019, n = 40) was superior to those achieved by state-of-the-art algorithms. While further optimization and validation of the models is recommended, automated data extraction expands the use application of (μ)CT imaging, while reducing subjectivity and workload. In conclusion, the proposed method offers a pipeline to efficiently monitor disease progression in animals and humans, allowing uniform translation from preclinical to clinical studies.

Mixed-UNet: Refined class activation mapping for weakly-supervised semantic segmentation with multi-scale inference

Deep learning techniques have shown great potential in medical image processing, particularly through accurate and reliable image segmentation on magnetic resonance imaging (MRI) scans or computed tomography (CT) scans, which allow the localization and diagnosis of lesions. However, training these segmentation models requires a large number of manually annotated pixel-level labels, which are time-consuming and labor-intensive, in contrast to image-level labels that are easier to obtain. It is imperative to resolve this problem through weakly-supervised semantic segmentation models using image-level labels as supervision since it can significantly reduce human annotation efforts. Most of the advanced solutions exploit class activation mapping (CAM). However, the original CAMs rarely capture the precise boundaries of lesions. In this study, we propose the strategy of multi-scale inference to refine CAMs by reducing the detail loss in single-scale reasoning. For segmentation, we develop a novel model named Mixed-UNet, which has two parallel branches in the decoding phase. The results can be obtained after fusing the extracted features from two branches. We evaluate the designed Mixed-UNet against several prevalent deep learning-based segmentation approaches on our dataset collected from the local hospital and public datasets. The validation results demonstrate that our model surpasses available methods under the same supervision level in the segmentation of various lesions from brain imaging.

Leveraging deep learning-based segmentation and contours-driven deformable registration for dose accumulation in abdominal structures

Discrepancies between planned and delivered dose to GI structures during radiation therapy (RT) of liver cancer may hamper the prediction of treatment outcomes. The purpose of this study is to develop a streamlined workflow for dose accumulation in a treatment planning system (TPS) during liver image-guided RT and to assess its accuracy when using different deformable image registration (DIR) algorithms. Fifty-six patients with primary and metastatic liver cancer treated with external beam radiotherapy guided by daily CT-on-rails (CTOR) were retrospectively analyzed. The liver, stomach and duodenum contours were auto-segmented on all planning CTs and daily CTORs using deep-learning methods. Dose accumulation was performed for each patient using scripting functionalities of the TPS and considering three available DIR algorithms based on: (i) image intensities only; (ii) intensities + contours; (iii) a biomechanical model (contours only). Planned and accumulated doses were converted to equivalent dose in 2Gy (EQD2) and normal tissue complication probabilities (NTCP) were calculated for the stomach and duodenum. Dosimetric indexes for the normal liver, GTV, stomach and duodenum and the NTCP values were exported from the TPS for analysis of the discrepancies between planned and the different accumulated doses. Deep learning segmentation of the stomach and duodenum enabled considerable acceleration of the dose accumulation process for the 56 patients. Differences between accumulated and planned doses were analyzed considering the 3 DIR methods. For the normal liver, stomach and duodenum, the distribution of the 56 differences in maximum doses (D2%) presented a significantly higher variance when a contour-driven DIR method was used instead of the intensity only-based method. Comparing the two contour-driven DIR methods, differences in accumulated minimum doses (D98%) in the GTV were >2Gy for 15 (27%) of the patients. Considering accumulated dose instead of planned dose in standard NTCP models of the duodenum demonstrated a high sensitivity of the duodenum toxicity risk to these dose discrepancies, whereas smaller variations were observed for the stomach. This study demonstrated a successful implementation of an automatic workflow for dose accumulation during liver cancer RT in a commercial TPS. The use of contour-driven DIR methods led to larger discrepancies between planned and accumulated doses in comparison to using an intensity only based DIR method, suggesting a better capability of these approaches in estimating complex deformations of the GI organs.

Cx22: A new publicly available dataset for deep learning-based segmentation of cervical cytology images

The segmentation of cervical cytology images plays an important role in the automatic analysis of cervical cytology screening. Although deep learning-based segmentation methods are well-developed in other image segmentation areas, their application in the segmentation of cervical cytology images is still in the early stage. The most important reason for the slow progress is the lack of publicly available and high-quality datasets, and the study on the deep learning-based segmentation methods may be hampered by the present datasets which are either artificial or plagued by the issue of false-negative objects. In this paper, we develop a new dataset of cervical cytology images named Cx22, which consists of the completely annotated labels of the cellular instances based on the open-source images released by our institute previously. Firstly, we meticulously delineate the contours of 14,946 cellular instances in1320 images that are generated by our proposed ROI-based label cropping algorithm. Then, we propose the baseline methods for the deep learning-based semantic and instance segmentation tasks based on Cx22. Finally, through the experiments, we validate the task suitability of Cx22, and the results reveal the impact of false-negative objects on the performance of the baseline methods. Based on our work, Cx22 can provide a foundation for fellow researchers to develop high-performance deep learning-based methods for the segmentation of cervical cytology images.Other detailed information and step-by-step guidance on accessing the dataset are made available to fellow researchers at https://github.com/LGQ330/Cx22.

Angiotensin-converting Enzyme-2 (ACE2) Expression in Pediatric Liver Disease.

The membrane protein angiotensin-converting enzyme-2 (ACE2) has gained notoriety as the receptor for severe acute respiratory syndrome coronavirus 2. Prior evidence has shown ACE2 is expressed within the liver but its function has not been fully discerned. Here, we utilized novel methodology to assess ACE2 expression in pediatric immune-mediated liver disease to better understand its presence in liver diseases and its role during infections such as COVID-19. We stained liver tissue with ACE2-specific immunofluorescent antibodies, analyzed via confocal microscopy. Computational deep learning-based segmentation models identified nuclei and cells, allowing the quantification of mean cellular and cytosolic immunofluorescent. Spatial transcriptomics provided high-throughput gene expression analysis in tissue to determine cellular composition for ACE2 expression. ACE2 plasma expression was quantified via enzyme-linked immunosorbent assay. High ACE2 expression was seen at the apical surface of cholangiocytes, with lower expression within hepatocyte cytosol and nonparenchymal cells ( P <0.001). Children with liver disease had higher ACE2 hepatic expression than pediatric control tissue ( P <0.001). Adult control tissue had higher expression than pediatric control ( P <0.001). Plasma ACE2 was not found to be statistically different between samples. Spatial transcriptomics identified cell composition of ACE2-expressing spots containing antibody-secreting cells. Our results show ACE2 expression throughout the liver, with strongest localization to cholangiocyte membranes. Machine learning can be used to rapidly identify hepatic cellular components for histologic analysis. ACE2 expression in the liver may be increased in pediatric liver disease. Future work is needed to better understand the role of ACE2 in chronic disease and acute infections.

Deep learning-based segmentation and classification of leaf images for detection of tomato plant disease.

Plants contribute significantly to the global food supply. Various Plant diseases can result in production losses, which can be avoided by maintaining vigilance. However, manually monitoring plant diseases by agriculture experts and botanists is time-consuming, challenging and error-prone. To reduce the risk of disease severity, machine vision technology (i.e., artificial intelligence) can play a significant role. In the alternative method, the severity of the disease can be diminished through computer technologies and the cooperation of humans. These methods can also eliminate the disadvantages of manual observation. In this work, we proposed a solution to detect tomato plant disease using a deep leaning-based system utilizing the plant leaves image data. We utilized an architecture for deep learning based on a recently developed convolutional neural network that is trained over 18,161 segmented and non-segmented tomato leaf images—using a supervised learning approach to detect and recognize various tomato diseases using the Inception Net model in the research work. For the detection and segmentation of disease-affected regions, two state-of-the-art semantic segmentation models, i.e., U-Net and Modified U-Net, are utilized in this work. The plant leaf pixels are binary and classified by the model as Region of Interest (ROI) and background. There is also an examination of the presentation of binary arrangement (healthy and diseased leaves), six-level classification (healthy and other ailing leaf groups), and ten-level classification (healthy and other types of ailing leaves) models. The Modified U-net segmentation model outperforms the simple U-net segmentation model by 98.66 percent, 98.5 IoU score, and 98.73 percent on the dice. InceptionNet1 achieves 99.95% accuracy for binary classification problems and 99.12% for classifying six segmented class images; InceptionNet outperformed the Modified U-net model to achieve higher accuracy. The experimental results of our proposed method for classifying plant diseases demonstrate that it outperforms the methods currently available in the literature.

Robustness of radiomics to variations in segmentation methods in multimodal brain MRI

Radiomics in neuroimaging uses fully automatic segmentation to delineate the anatomical areas for which radiomic features are computed. However, differences among these segmentation methods affect radiomic features to an unknown extent. A scan-rescan dataset (n = 46) of T1-weighted and diffusion tensor images was used. Subjects were split into a sleep-deprivation and a control group. Scans were segmented using four segmentation methods from which radiomic features were computed. First, we measured segmentation agreement using the Dice-coefficient. Second, robustness and reproducibility of radiomic features were measured using the intraclass correlation coefficient (ICC). Last, difference in predictive power was assessed using the Friedman-test on performance in a radiomics-based sleep deprivation classification application. Segmentation agreement was generally high (interquartile range = 0.77–0.90) and median feature robustness to segmentation method variation was higher (ICC > 0.7) than scan-rescan reproducibility (ICC 0.3–0.8). However, classification performance differed significantly among segmentation methods (p < 0.001) ranging from 77 to 84%. Accuracy was higher for more recent deep learning-based segmentation methods. Despite high agreement among segmentation methods, subtle differences significantly affected radiomic features and their predictive power. Consequently, the effect of differences in segmentation methods should be taken into account when designing and evaluating radiomics-based research methods.

Endolymphatic space is age-dependent

Knowledge of the physiological endolymphatic space (ELS) is necessary to estimate endolymphatic hydrops (ELH) in patients with vestibulocochlear syndromes. Therefore, the current study investigated age-dependent changes in the ELS of participants with normal vestibulocochlear testing. Sixty-four ears of 32 participants with normal vestibulocochlear testing aged between 21 and 75 years (45.8 ± 17.2 years, 20 females, 30 right-handed, two left-handed) were examined by intravenous delayed gadolinium-enhanced magnetic resonance imaging of the inner ear (iMRI). Clinical diagnostics included neuro-otological assessment, video-oculography during caloric stimulation, and head-impulse test. iMRI data analysis provided semi-quantitative visual grading and automatic algorithmic quantitative segmentation of ELS volume (3D, mm3) using a deep learning-based segmentation of the inner ear’s total fluid space (TFS) and volumetric local thresholding, as described earlier. As a result, following a 4-point ordinal scale, a mild ELH (grade 1) was found in 21/64 (32.8%) ears uni- or bilaterally in either cochlear, vestibulum, or both. Age and ELS were found to be positively correlated for the inner ear (r(64) = 0.33, p < 0.01), and vestibulum (r(64) = 0.25, p < 0.05). For the cochlea, the values correlated positively without reaching significance (r(64) = 0.21). In conclusion, age-dependent increases of the ELS should be considered when evaluating potential ELH in single subjects and statistical group comparisons.

Two-step machine learning to diagnose and predict involvement of lungs in COVID-19 and pneumonia using CT radiomics

ObjectiveTo develop a two-step machine learning (ML) based model to diagnose and predict involvement of lungs in COVID-19 and non COVID-19 pneumonia patients using CT chest radiomic features. MethodsThree hundred CT scans (3-classes: 100 COVID-19, 100 pneumonia, and 100 healthy subjects) were enrolled in this study. Diagnostic task included 3-class classification. Severity prediction score for COVID-19 and pneumonia was considered as mild (0-25%), moderate (26-50%), and severe (>50%). Whole lungs were segmented utilizing deep learning-based segmentation. Altogether, 107 features including shape, first-order histogram, second and high order texture features were extracted. Pearson correlation coefficient (PCC≥90%) followed by different features selection algorithms were employed. ML-based supervised algorithms (Naïve Bays, Support Vector Machine, Bagging, Random Forest, K-nearest neighbors, Decision Tree and Ensemble Meta voting) were utilized. The optimal model was selected based on precision, recall and area-under-curve (AUC) by randomizing the training/validation, followed by testing using the test set. ResultsNine pertinent features (2 shape, 1 first-order, and 6 second-order) were obtained after features selection for both phases. In diagnostic task, the performance of 3-class classification using Random Forest was 0.909±0.026, 0.907±0.056, 0.902±0.044, 0.939±0.031, and 0.982±0.010 for precision, recall, F1-score, accuracy, and AUC, respectively. The severity prediction task using Random Forest achieved 0.868±0.123 precision, 0.865±0.121 recall, 0.853±0.139 F1-score, 0.934±0.024 accuracy, and 0.969±0.022 AUC. ConclusionThe two-phase ML-based model accurately classified COVID-19 and pneumonia patients using CT radiomics, and adequately predicted severity of lungs involvement. This 2-steps model showed great potential in assessing COVID-19 CT images towards improved management of patients.

Two-Stage CNN Whole Heart Segmentation Combining Image Enhanced Attention Mechanism and Metric Classification.

Accurate segmentation of multiple tissues and organs in cardiac medical imaging is of great value in computer-aided cardiovascular diagnosis. However, it is challenging due to the complex distribution of various tissues and organs in cardiac MRI (magnetic resonance imaging) slices, low discriminative and large spanning organs. To handle these problems, a two-stage CNN (convolutional neural network) segmentation method based on the combination of Log-Gabor filter attention mechanism and metric classification is proposed. The Log-Gabor filterbank is applied to selectively enhance the texture information and contour information of each tissue and organ, and the spatial and channel attention mechanism jointly with the varying kernel size of Log-Gabor filterbank is incorporated into the codec structure to adaptively extract target features of different sizes and focus on the discriminative features in the network. To solve the problem of insufficient segmentation on subtle and adherent edges involving different tissues, a metric classification network is incorporated to finely optimize the hard-to-be-segmented boundaries. The proposed method was tested on cardiac MRI data set to segment 7 cardiac tissues, and the rationality and effectiveness of the method were verified. In comparison to a series of deep learning-based segmentation models, the proposed method achieves competitive performance.

Multiscale depthwise separable convolution based network for high-resolution image segmentation

ABSTRACT Deep learning-based segmentation methods have demonstrated significant performance over their traditional counterparts. However, striving for better accuracy with such networks usually leads to the deterioration of the network’s computational efficiency, thereby rendering them inefficient for deployment on resource constraint devices. Establishing the required tradeoff between the accuracy of pixel prediction and computational efficiency remains challenging. In this article, a lightweight multiscale segmentation framework is proposed. We leverage the representation power of different receptive fields to attain optimal accuracy while maintaining computational efficiency by embedding the sparse network architecture with the depthwise separable convolution at the multiscale level. Experimental results from two challenging remote sensing segmentation datasets show that the proposed network can achieve substantial pixel prediction accuracy at relatively low computational overhead compared to state-of-the-art networks.