Computer-aided Diagnostic System Research Articles

Architectural Distortion-Based Digital Mammograms Classification Using Depth Wise Convolutional Neural Network.

Simple SummaryBreast cancer is leading cancer increases the death rate in women. Early diagnosis of breast cancer in women can save their lives. The current study proposed a novel scheme to detect architectural distortion from mammogram images to predict breast cancer using a deep learning approach. Results are evaluated on a public and a private dataset which may help to improve the diagnostic ability of breast cancer of radiologists and doctors in daily clinical routines. Furthermore, the proposed method achieved maximum accuracy as compared with previous approaches. This study can be interesting and valuable in the healthcare predictive modeling domain and will add a real contribution to society.Architectural distortion is the third most suspicious appearance on a mammogram representing abnormal regions. Architectural distortion (AD) detection from mammograms is challenging due to its subtle and varying asymmetry on breast mass and small size. Automatic detection of abnormal ADs regions in mammograms using computer algorithms at initial stages could help radiologists and doctors. The architectural distortion star shapes ROIs detection, noise removal, and object location, affecting the classification performance, reducing accuracy. The computer vision-based technique automatically removes the noise and detects the location of objects from varying patterns. The current study investigated the gap to detect architectural distortion ROIs (region of interest) from mammograms using computer vision techniques. Proposed an automated computer-aided diagnostic system based on architectural distortion using computer vision and deep learning to predict breast cancer from digital mammograms. The proposed mammogram classification framework pertains to four steps such as image preprocessing, augmentation and image pixel-wise segmentation. Architectural distortion ROI’s detection, training deep learning, and machine learning networks to classify AD’s ROIs into malignant and benign classes. The proposed method has been evaluated on three databases, the PINUM, the CBIS-DDSM, and the DDSM mammogram images, using computer vision and depth-wise 2D V-net 64 convolutional neural networks and achieved 0.95, 0.97, and 0.98 accuracies, respectively. Experimental results reveal that our proposed method outperforms as compared with the ShuffelNet, MobileNet, SVM, K-NN, RF, and previous studies.

TNSNet: Thyroid nodule segmentation in ultrasound imaging using soft shape supervision

Background and objectives: Thyroid nodules are a common disorder of the endocrine system. Segmentation of thyroid nodules on ultrasound images is an important step in the evaluation and diagnosis of nodules and an initial step in computer-aided diagnostic systems. The accuracy and consistency of segmentation remain a challenge due to the low contrast, speckle noise, and low resolution of ultrasound images. Therefore, the study of deep learning-based algorithms for thyroid nodule segmentation is important. This study utilizes soft shape supervision to improve the performance of detection and segmentation of boundaries of nodules. Soft shape supervision can emphasize the boundary features and assist the network in segmenting nodules accurately. Methods: We propose a dual-path convolution neural network, including region and shape paths, which use DeepLabV3+ as the backbone. Soft shape supervision blocks are inserted between the two paths to implement cross-path attention mechanisms. The blocks enhance the representation of shape features and add them to the region path as auxiliary information. Thus, the network can accurately detect and segment thyroid nodules. Results: We collect 3786 ultrasound images of thyroid nodules to train and test our network. Compared with the ground truth, the test results achieve an accuracy of 95.81% and a DSC of 85.33. The visualization results also suggest that the network has learned clear and accurate boundaries of the nodules. The evaluation metrics and visualization results demonstrate the superior segmentation performance of the network to other classical deep learning-based networks. Conclusions: The proposed dual-path network can accurately realize automatic segmentation of thyroid nodules on ultrasound images. It can also be used as an initial step in computer-aided diagnosis. It shows superior performance to other classical methods and demonstrates the potential for accurate segmentation of nodules in clinical applications.

Role of Optical Coherence Tomography Imaging in Predicting Progression of Age-Related Macular Disease: A Survey.

In developed countries, age-related macular degeneration (AMD), a retinal disease, is the main cause of vision loss in the elderly. Optical Coherence Tomography (OCT) is currently the gold standard for assessing individuals for initial AMD diagnosis. In this paper, we look at how OCT imaging can be used to diagnose AMD. Our main aim is to examine and compare automated computer-aided diagnostic (CAD) systems for diagnosing and grading of AMD. We provide a brief summary, outlining the main aspects of performance assessment and providing a basis for current research in AMD diagnosis. As a result, the only viable alternative is to prevent AMD and stop both this devastating eye condition and unwanted visual impairment. On the other hand, the grading of AMD is very important in order to detect early AMD and prevent patients from reaching advanced AMD disease. In light of this, we explore the remaining issues with automated systems for AMD detection based on OCT imaging, as well as potential directions for diagnosis and monitoring systems based on OCT imaging and telemedicine applications.

Explainable COVID-19 Detection on Chest X-rays Using an End-to-End Deep Convolutional Neural Network Architecture

The coronavirus pandemic is spreading around the world. Medical imaging modalities such as radiography play an important role in the fight against COVID-19. Deep learning (DL) techniques have been able to improve medical imaging tools and help radiologists to make clinical decisions for the diagnosis, monitoring and prognosis of different diseases. Computer-Aided Diagnostic (CAD) systems can improve work efficiency by precisely delineating infections in chest X-ray (CXR) images, thus facilitating subsequent quantification. CAD can also help automate the scanning process and reshape the workflow with minimal patient contact, providing the best protection for imaging technicians. The objective of this study is to develop a deep learning algorithm to detect COVID-19, pneumonia and normal cases on CXR images. We propose two classifications problems, (i) a binary classification to classify COVID-19 and normal cases and (ii) a multiclass classification for COVID-19, pneumonia and normal. Nine datasets and more than 3200 COVID-19 CXR images are used to assess the efficiency of the proposed technique. The model is trained on a subset of the National Institute of Health (NIH) dataset using swish activation, thus improving the training accuracy to detect COVID-19 and other pneumonia. The models are tested on eight merged datasets and on individual test sets in order to confirm the degree of generalization of the proposed algorithms. An explainability algorithm is also developed to visually show the location of the lung-infected areas detected by the model. Moreover, we provide a detailed analysis of the misclassified images. The obtained results achieve high performances with an Area Under Curve (AUC) of 0.97 for multi-class classification (COVID-19 vs. other pneumonia vs. normal) and 0.98 for the binary model (COVID-19 vs. normal). The average sensitivity and specificity are 0.97 and 0.98, respectively. The sensitivity of the COVID-19 class achieves 0.99. The results outperformed the comparable state-of-the-art models for the detection of COVID-19 on CXR images. The explainability model shows that our model is able to efficiently identify the signs of COVID-19.

An efficient real-time colonic polyp detection with YOLO algorithms trained by using negative samples and large datasets

Colorectal cancer (CRC) is one of the common types of cancer with a high mortality rate. Colonoscopy is the gold standard for CRC screening and significantly reduces CRC mortality. However, due to many factors, the rate of missed polyps, which are the precursors of colorectal cancer, is high in practice. Therefore, many artificial intelligence-based computer-aided diagnostic systems have been presented to increase the detection rate of missed polyps. In this article, we present deep learning-based methods for reliable computer-assisted polyp detection. The proposed methods differ from state-of-the-art methods as follows. First, we improved the performances of YOLOv3 and YOLOv4 object detection algorithms by integrating Cross Stage Partial Network (CSPNet) for real-time and high-performance automatic polyp detection. Then, we utilized advanced data augmentation techniques and transfer learning to improve the performance of polyp detection. Next, for further improving the performance of polyp detection using negative samples, we substituted the Sigmoid-weighted Linear Unit (SiLU) activation functions instead of the Leaky ReLU and Mish activation functions, and Complete Intersection over Union (CIoU) as the loss function. In addition, we present a comparative analysis of these activation functions for polyp detection. We applied the proposed methods on the recently published novel datasets, which are the SUN polyp database and the PICCOLO database. Additionally, we investigated the proposed models for MICCAI Sub-Challenge on Automatic Polyp Detection in Colonoscopy dataset. The proposed methods outperformed the other studies in both real-time performance and polyp detection accuracy.

Nuclei Segmentation on Histopathology Images of Breast Carcinoma.

With the use of computer-aided diagnostic systems, the automatic detection and segmentation of the cell nuclei have become essential in pathology due to cellular nuclei counting and nuclear pleomorphism analysis are critical for the classification and grading of breast cancer histopathology. This work describes a methodology for automatic detection and segmentation of cellular nuclei in breast cancer histopathology images obtained from the BreakHis database, the Standford tissue microarray database, and the Breast Cancer Cell Segmentation database. The proposed scheme is based on the characterization of Hematoxylin and Eosin (H&E) staining, size, and shape features. In addition, we use the information obtained from morphological transformations and adaptive intensity adjustments to detect and separate each cell nucleus detected in the image. The segmentation was carried out by testing the proposed methodology in a histological breast cancer database that provides the associated groundtruth segmentation. Subsequently, the Sørensen-Dice similarity coefficient was calculated to analyze the suitability of the results.Clinical relevance- In this work, the detection and segmentation of cell nuclei in breast cancer histological images are carried out automatically. The method can identify cell nuclei regardless of variations in the level of staining and image magnification. Moreover, a granulometric analysis of the components allows identifying cell clumps and segment them into individual cell nuclei. Improved identification of cell nuclei under different image conditions was demonstrated to reach a sensitivity average of 0.76 ± 0.12. The results provide a base for further and complex processes such as cell counting, feature analysis, and nuclear pleomorphism, which are relevant tasks in the evaluation and diagnostic performed by the expert pathologist.

Automated COVID-19 and Heart Failure Detection Using DNA Pattern Technique with Cough Sounds.

COVID-19 and heart failure (HF) are common disorders and although they share some similar symptoms, they require different treatments. Accurate diagnosis of these disorders is crucial for disease management, including patient isolation to curb infection spread of COVID-19. In this work, we aim to develop a computer-aided diagnostic system that can accurately differentiate these three classes (normal, COVID-19 and HF) using cough sounds. A novel handcrafted model was used to classify COVID-19 vs. healthy (Case 1), HF vs. healthy (Case 2) and COVID-19 vs. HF vs. healthy (Case 3) automatically using deoxyribonucleic acid (DNA) patterns. The model was developed using the cough sounds collected from 241 COVID-19 patients, 244 HF patients, and 247 healthy subjects using a hand phone. To the best our knowledge, this is the first work to automatically classify healthy subjects, HF and COVID-19 patients using cough sounds signals. Our proposed model comprises a graph-based local feature generator (DNA pattern), an iterative maximum relevance minimum redundancy (ImRMR) iterative feature selector, with classification using the k-nearest neighbor classifier. Our proposed model attained an accuracy of 100.0%, 99.38%, and 99.49% for Case 1, Case 2, and Case 3, respectively. The developed system is completely automated and economical, and can be utilized to accurately detect COVID-19 versus HF using cough sounds.

A Deep Learning Pipeline for Grade Groups Classification Using Digitized Prostate Biopsy Specimens.

Prostate cancer is a significant cause of morbidity and mortality in the USA. In this paper, we develop a computer-aided diagnostic (CAD) system for automated grade groups (GG) classification using digitized prostate biopsy specimens (PBSs). Our CAD system aims to firstly classify the Gleason pattern (GP), and then identifies the Gleason score (GS) and GG. The GP classification pipeline is based on a pyramidal deep learning system that utilizes three convolution neural networks (CNN) to produce both patch- and pixel-wise classifications. The analysis starts with sequential preprocessing steps that include a histogram equalization step to adjust intensity values, followed by a PBSs’ edge enhancement. The digitized PBSs are then divided into overlapping patches with the three sizes: 100 × 100 (), 150 × 150 (), and 200 × 200 (), pixels, and 75% overlap. Those three sizes of patches represent the three pyramidal levels. This pyramidal technique allows us to extract rich information, such as that the larger patches give more global information, while the small patches provide local details. After that, the patch-wise technique assigns each overlapped patch a label as GP categories (1 to 5). Then, the majority voting is the core approach for getting the pixel-wise classification that is used to get a single label for each overlapped pixel. The results after applying those techniques are three images of the same size as the original, and each pixel has a single label. We utilized the majority voting technique again on those three images to obtain only one. The proposed framework is trained, validated, and tested on 608 whole slide images (WSIs) of the digitized PBSs. The overall diagnostic accuracy is evaluated using several metrics: precision, recall, F1-score, accuracy, macro-averaged, and weighted-averaged. The () has the best accuracy results for patch classification among the three CNNs, and its classification accuracy is 0.76. The macro-averaged and weighted-average metrics are found to be around 0.70–0.77. For GG, our CAD results are about 80% for precision, and between 60% to 80% for recall and F1-score, respectively. Also, it is around 94% for accuracy and NPV. To highlight our CAD systems’ results, we used the standard ResNet50 and VGG-16 to compare our CNN’s patch-wise classification results. As well, we compared the GG’s results with that of the previous work.

Ensemble based machine learning approach for prediction of glioma and multi-grade classification

Glioma is the most pernicious cancer of the nervous system, with histological grade influencing the survival of patients. Despite many studies on the multimodal treatment approach, survival time remains brief. In this study, a novel two-stage ensemble of an ensemble-type machine learning-based predictive framework for glioma detection and its histograde classification is proposed. In the proposed framework, five characteristics belonging to 135 subjects were considered: human telomerase reverse transcriptase (hTERT), chitinase-like protein (YKL-40), interleukin 6 (IL-6), tissue inhibitor of metalloproteinase-1 (TIMP-1) and neutrophil/lymphocyte ratio (NLR). These characteristics were examined using distinctive ensemble-based machine learning classifiers and combination strategies to develop a computer-aided diagnostic system for the non-invasive prediction of glioma cases and their grade. In the first stage, the analysis was conducted to classify glioma cases and control subjects. Machine learning approaches were applied in the second stage to classify the recognised glioma cases into three grades, from grade II, which has a good prognosis, to grade IV, which is also known as glioblastoma. All experiments were evaluated with a five-fold cross-validation method, and the classification results were analysed using different statistical parameters. The proposed approach obtained a high value of accuracy and other statistical parameters compared with other state-of-the-art machine learning classifiers. Therefore, the proposed framework can be utilised for designing other intervention strategies for the prediction of glioma cases and their grades.

Automatic glottis segmentation for laryngeal endoscopic images based on U-Net

The glottis’s morphology not only reflects vocal and respiratory information, but also plays an important role in the diagnosis of laryngeal diseases. The glottis segmentation is a primary step in computer-aided diagnostic system, however is challenging due to various shapes of glottis, low contrast with surrounding tissues, the existence of laryngeal diseases and so on. In this paper, a deep attention network based on U-Net with color normalization operation (CN-DA-Unet) is proposed to achieve an end-to-end segmentation of the glottal area for the first time. The original images are first processed by color normalization to reduce the adverse effects of low contrast and large differences in colors between different images. The normalized images are then sent to the proposed DA-Unet for feature extraction. In this network, residual structure is incorporated to extract rich features from deep neural networks. After extracting features, a feature pyramid attention (FPA) module is applied to enhance the semantic information of the glottal area. These features are up-sampled and added to the features from the corresponding encoding layer for several times to obtain the final segmented image. The proposed approach is tested on laryngeal images of an in–house dataset including images from healthy subjects and pathologic subjects. Its performance is evaluated by several reliable and popular evaluation metrics, achieving the dice coefficient of 92.9%, sensitivity of 93.5% and precision of 92.6%. These results demonstrate the effectiveness of our proposed approach and the better performance comparing with several popular networks.

A Novel 3-D-CAD for Breast and Lesion Segmentation of Axial Breast DCE-MRI

Abstract The present paper proposes a fully automated three-dimensional (3-D) system for breast and lesion segmentation of Dynamic Contrast Enhanced MRI (DCE-MRI). Such a system as the Computer-Aided Diagnostic system (CAD) can be used to support radiologists by marking suspicious areas. The proposed 3-D-CAD system includes three modules. The first one concerns breast area segmentation based on image content analysis, the Moore-Neighbor tracing algorithm, and the Dijkstra algorithm. The second one concerns the automation of locating and selecting lesions that starts by preprocessing the already segmented breast regions; then, a K-means algorithm allows extraction of regions including suspicious tissues. The third one is the superimposition of all detected lesions from selected slices to create a 3-D view of the lesion. The 3-D reconstruction is based on the Marching Cube algorithm. The validation of breast area segmentation reveals the robustness of the proposed process versus different breast densities, complex forms, and challenging cases. The segmentation of the breast part from 120 slices with the proposed method is achieved in 20.57 ± 5.2 s, which is faster than existing methods. In addition to calculated metrics as Dice similarity coefficient (DSC) and rates of true positives, the module of lesion extraction is validated by experimented radiologists. The proposed workflow of this module shows competitive results compared to the existing methods of automated lesion segmentation and a total ability for eliminating the extra regions to lesions.

Feature based CT image registration of liver cancer.

Computer-aided diagnostic systems (CADS) assist radiologists in classifying liver cancer using computed tomography (CT) images. To enhance diagnosis performance, image sequences are recorded at various time points in a single/multi-view format. Mutual information (MI) is a widely used medical image registration metric with a high rate of success, but it can result in misregistration due to a lack of spatial details. To address this issue and to establish anatomical correspondence between multi-phase CT images of the liver, a features-based technique is developed in this article. The proposed model uses fixed and moving images as inputs, with both images having the same dimensions. The registered images are the two images that differ in terms of their combinations/colors. In the output registered images, the tumor in the liver portion has classes with viewpoints. There is an appropriate way to view the tumor, and the output registered images should permit concluding that the registered image of the delayed phases, with a longer delay time, contains the most region portion within the output registered image. The detected and matched values are greater than the values of the feature outcomes. Having a large tumor provides valuable information in the presenting form for discussing the variation of the various phases and delayed testing results. And this will aid the radiologist in making an accurate diagnosis.

Case-Base Maintenance: An Approach Based on Active Semi-Supervised Learning

Case-Base Maintenance (CBM) becomes of great importance when implementing a Computer-Aided Diagnostic (CAD) system using Case-Based Reasoning (CBR). Since it is essential for the learning to avoid the case-base degradation, this work aims to build and maintain a quality case base while overcoming the difficulty of assembling labeled case bases, traditionally assumed to exist or determined by human experts. The proposed approach takes advantage of large volumes of unlabeled data to select valuable cases to add to the case base while monitoring retention to avoid performance degradation and to build a compact quality case base. We use machine learning techniques to cope with this challenge: an Active Semi-Supervised Learning approach is proposed to overcome the bottleneck of scarcity of labeled data. In order to acquire a quality case base, we target its performance criterion. Case selection and retention are assessed according to three combined sampling criteria: informativeness, representativeness, and diversity. We support our approach with empirical evaluations using different benchmark data sets. Based on experimentation, the proposed approach achieves good classification accuracy with a small number of retained cases, using a small training set as a case base.

USE OF DEEP LEARNING TO DIAGNOSE COVID-19 BASED ON COMPUTED TOMOGRAPHY IMAGES

The newly identified Coronavirus pneumonia, later called COVID-19, is highly transmissible and pathogenic. The most common symptoms of this disease are dry cough, sore throat, and fever. Symptoms can progress to a severe form of pneumonia with critical complications, including septic shock, pulmonary edema, acute respiratory distress syndrome, and multiple organ failure. A major obstacle in controlling the spread of this disease is the inefficiency and scarcity of medical tests. Increasing efforts have been made to develop deep learning (DL) methods to diagnose COVID-19 based on tomography images. These computer-aided diagnostic systems can assist in the early detection of abnormalities in COVID-19 and facilitate the monitoring of disease progression, potentially reducing mortality rates. In this study, we compared the popular resource extraction structures based on deep learning for the automatic classification of COVID-19. To obtain a more precise method, which is an essential learning component, a set of deep convolutional neural networks (CNN) was chosen to train our model. The performance of the proposed method was validated using a COVID-19 dataset with computed tomography (CT) images. This dataset is available to the public and contains hundreds of positive CT scans for the disease. DL methods were performed and the best classified CNN was able to achieve excellent diagnostic results for COVID-19.

Lung Cancer Detection and Improving Accuracy Using Linear Subspace Image Classification Algorithm.

The ability to identify lung cancer at an early stage is critical, because it can help patients live longer. However, predicting the affected area while diagnosing cancer is a huge challenge. An intelligent computer-aided diagnostic system can be utilized to detect and diagnose lung cancer by detecting the damaged region. The suggested Linear Subspace Image Classification Algorithm (LSICA) approach classifies images in a linear subspace. This methodology is used to accurately identify the damaged region, and it involves three steps: image enhancement, segmentation, and classification. The spatial image clustering technique is used to quickly segment and identify the impacted area in the image. LSICA is utilized to determine the accuracy value of the affected region for classification purposes. Therefore, a lung cancer detection system with classification-dependent image processing is used for lung cancer CT imaging. Therefore, a new method to overcome these deficiencies of the process for detection using LSICA is proposed in this work on lung cancer. MATLAB has been used in all programs. A proposed system designed to easily identify the affected region with help of the classification technique to enhance and get more accurate results.

DeepCervix: A deep learning-based framework for the classification of cervical cells using hybrid deep feature fusion techniques

Cervical cancer, one of the most common fatal cancers among women, can be prevented by regular screening to detect any precancerous lesions at early stages and treat them. Pap smear test is a widely performed screening technique for early detection of cervical cancer, whereas this manual screening method suffers from high false-positive results because of human errors. To improve the manual screening practice, machine learning (ML) and deep learning (DL) based computer-aided diagnostic (CAD) systems have been investigated widely to classify cervical Pap cells. Most of the existing studies require pre-segmented images to obtain good classification results. In contrast, accurate cervical cell segmentation is challenging because of cell clustering. Some studies rely on handcrafted features, which cannot guarantee the classification stage's optimality. Moreover, DL provides poor performance for a multiclass classification task when there is an uneven distribution of data, which is prevalent in the cervical cell dataset. This investigation has addressed those limitations by proposing DeepCervix, a hybrid deep feature fusion (HDFF) technique based on DL, to classify the cervical cells accurately. Our proposed method uses various DL models to capture more potential information to enhance classification performance. Our proposed HDFF method is tested on the publicly available SIPaKMeD dataset and compared the performance with base DL models and the late fusion (LF) method. For the SIPaKMeD dataset, we have obtained the state-of-the-art classification accuracy of 99.85%, 99.38%, and 99.14% for 2-class, 3-class, and 5-class classification. This method is also tested on the Herlev dataset and achieves an accuracy of 98.32% for 2-class and 90.32% for 7-class classification. The source code of the DeepCervix model is available at: https://github.com/Mamunur-20/DeepCervix.

Multi-channel Chan-Vese model for unsupervised segmentation of nuclei from breast histopathological images

T he pathologist determines the malignancy of a breast tumor by studying the histopathological images. In particular, the characteristics and distribution of nuclei contribute greatly to the decision process. Hence, the segmentation of nuclei constitutes a crucial task in the classification of breast histopathological images. Manual analysis of these images is subjective, tedious and susceptible to human error. Consequently, the development of computer-aided diagnostic systems for analysing these images have become a vital factor in the domain of medical imaging. However, the usage of medical image processing techniques to segment nuclei is challenging due to the diverse structure of the cells, poor staining process, the occurrence of artifacts, etc. Although supervised computer-aided systems for nuclei segmentation is popular, it is dependent on the availability of standard annotated datasets. In this regard, this work presents an unsupervised method based on Chan-Vese model to segment nuclei from breast histopathological images. The proposed model utilizes multi-channel color information to efficiently segment the nuclei. Also, this study proposes a pre-processing step to select appropriate color channel such that it discriminates nuclei from the background region. An extensive evaluation of the proposed model on two challenging datasets demonstrates its validity and effectiveness.

Deep learning for abdominal ultrasound: A computer-aided diagnostic system for the severity of fatty liver.

The prevalence of nonalcoholic fatty liver disease is increasing over time worldwide, with similar trends to those of diabetes and obesity. A liver biopsy, the gold standard of diagnosis, is not favored due to its invasiveness. Meanwhile, noninvasive evaluation methods of fatty liver are still either very expensive or demonstrate poor diagnostic performances, thus, limiting their applications. We developed neural network-based models to assess fatty liver and classify the severity using B-mode ultrasound (US) images. We followed standards for reporting of diagnostic accuracy guidelines to report this study. In this retrospective study, we utilized B-mode US images from a consecutive series of patients to develop four-class, two-class, and three-class diagnostic prediction models. The images were eligible if confirmed by at least two gastroenterologists. We compared pretrained convolutional neural network models, consisting of visual geometry group (VGG)19, ResNet-50 v2, MobileNet v2, Xception, and Inception v2. For validation, we utilized 20% of the dataset resulting in >100 images for each severity category. There were 21,855 images from 2,070 patients classified as normal (N = 11,307), mild (N = 4,467), moderate (N = 3,155), or severe steatosis (N = 2,926). We used ResNet-50 v2 for the final model as the best ones. The areas under the receiver operating characteristic curves were 0.974 (mild steatosis vs others), 0.971 (moderate steatosis vs others), 0.981 (severe steatosis vs others), 0.985 (any severity vs normal), and 0.996 (moderate-to-severe steatosis/clinically abnormal vs normal-to-mild steatosis/clinically normal). Our deep learning models achieved comparable predictive performances to the most accurate, yet expensive, noninvasive diagnostic methods for fatty liver. Because of the discriminative ability, including for mild steatosis, significant impacts on clinical applications for fatty liver are expected. However, we need to overcome machine-dependent variation, motion artifacts, lacking of second confirmation from any other tools, and hospital-dependent regional bias.

Attention mechanism and mixup data augmentation for classification of COVID-19 Computed Tomography images

The Coronavirus disease is quickly spreading all over the world and the emergency situation is still out of control. Latest achievements of deep learning algorithms suggest the use of deep Convolutional Neural Network to implement a computer-aided diagnostic system for automatic classification of COVID-19 CT images. In this paper, we propose to employ a feature-wise attention layer in order to enhance the discriminative features obtained by convolutional networks. Moreover, the original performance of the network has been improved using the mixup data augmentation technique. This work compares the proposed attention-based model against the stacked attention networks, and traditional versus mixup data augmentation approaches. We deduced that feature-wise attention extension, while outperforming the stacked attention variants, achieves remarkable improvements over the baseline convolutional neural networks. That is, ResNet50 architecture extended with a feature-wise attention layer obtained 95.57% accuracy score, which, to best of our knowledge, fixes the state-of-the-art in the challenging COVID-CT dataset.

A novel weight pruning strategy for light weight neural networks with application to the diagnosis of skin disease

In recent years, deep learning-based models have achieved significant advances in manifold computer vision problems, but tedious parameter tuning has complicated their application to computer-aided diagnostic (CAD) systems. As such, this study introduces a novel pruning strategy to improve the accuracy of five lightweight deep convolutional neural network (DCNN) architectures applied to the classification of skin disease. Unlike conventional pruning methods (such as optimal brain surgeon), the proposed technique does not change the model size yet improves performance after fine tuning. This training approach, intended to improve accuracy without increasing model complexity, is experimentally verified using 1167 pathological images. The clinical data included 11 different skin disease types collected over the past ten years, with varying image quantities in each category. A novel hierarchical pruning method, based on standard deviation, is then developed and used to prune parameters in each convolution layer according to the different weight distributions. This training strategy achieves an 83.5% Top-1 accuracy using a pruned MnasNet (12.5 MB), which is 1.8% higher than that of unpruned InceptionV3 (256 MB). Comparative experiments using other networks (MobileNetV2, SqueezeNet, ShuffleNetV2, Xception, ResNet50, DenseNet121) and dataset (HAM10000) also demonstrate consistent improvements when adopting the proposed model training technique. This distinctive robustness across various network types and simple deployment demonstrates the potential of this methodology for generalization to other computer vision tasks.