An interpretable dual attention network for diabetic retinopathy grading: IDANet.

Diabetic retinopathy (DR) is the most prevalent cause of visual impairment in adults worldwide. Typically, patients with DR do not show symptoms until later stages, by which time it may be too late to receive effective treatment. DR Grading is challenging because of the small size and variation in lesion patterns. The key to fine-grained DR grading is to discover more discriminating elements such as cotton wool, hard exudates, hemorrhages, microaneurysms etc. Although deep learning models like convolutional neural networks (CNN) seem ideal for the automated detection of abnormalities in advanced clinical imaging, small-size lesions are very hard to distinguish by using traditional networks. This work proposes a bi-directional spatial and channel-wise parallel attention based network to learn discriminative features for diabetic retinopathy grading. The proposed attention block plugged with a backbone network helps to extract features specific to fine-grained DR-grading. This scheme boosts classification performance along with the detection of small-sized lesion parts. Extensive experiments are performed on four widely used benchmark datasets for DR grading, and performance is evaluated on different quality metrics. Also, for model interpretability, activation maps are generated using the LIME method to visualize the predicted lesion parts. In comparison with state-of-the-art methods, the proposed IDANet exhibits better performance for DR grading and lesion detection.

CoInNet: A Convolution-Involution Network With a Novel Statistical Attention for Automatic Polyp Segmentation.

Polyps are very common abnormalities in human gastrointestinal regions. Their early diagnosis may help in reducing the risk of colorectal cancer. Vision-based computer-aided diagnostic systems automatically identify polyp regions to assist surgeons in their removal. Due to their varying shape, color, size, texture, and unclear boundaries, polyp segmentation in images is a challenging problem. Existing deep learning segmentation models mostly rely on convolutional neural networks that have certain limitations in learning the diversity in visual patterns at different spatial locations. Further, they fail to capture inter-feature dependencies. Vision transformer models have also been deployed for polyp segmentation due to their powerful global feature extraction capabilities. But they too are supplemented by convolution layers for learning contextual local information. In the present paper, a polyp segmentation model CoInNet is proposed with a novel feature extraction mechanism that leverages the strengths of convolution and involution operations and learns to highlight polyp regions in images by considering the relationship between different feature maps through a statistical feature attention unit. To further aid the network in learning polyp boundaries, an anomaly boundary approximation module is introduced that uses recursively fed feature fusion to refine segmentation results. It is indeed remarkable that even tiny-sized polyps with only 0.01% of an image area can be precisely segmented by CoInNet. It is crucial for clinical applications, as small polyps can be easily overlooked even in the manual examination due to the voluminous size of wireless capsule endoscopy videos. CoInNet outperforms thirteen state-of-the-art methods on five benchmark polyp segmentation datasets.

GeneViT: Gene Vision Transformer with Improved DeepInsight for cancer classification.

Analysis of gene expression data is crucial for disease prognosis and diagnosis. Gene expression data has high redundancy and noise that brings challenges in extracting disease information. Over the past decade, several conventional machine learning and deep learning models have been developed for classification of diseases using gene expressions. In recent years, vision transformer networks have shown promising performance in many fields due to their powerful attention mechanism that provides a better insight into the data characteristics. However, these network models have not been explored for gene expression analysis. In this paper, a method for classifying cancerous gene expression is presented that uses a Vision transformer. The proposed method first performs dimensionality reduction using a stacked autoencoder followed by an Improved DeepInsight algorithm that converts the data into image format. The data is then fed to the vision transformer for building the classification model. Performance of the proposed classification model is evaluated on ten benchmark datasets having binary classes or multiple classes. Its performance is also compared with nine existing classification models. The experimental results demonstrate that the proposed model outperforms existing methods. The t-SNE plots demonstrate the distinctive feature learning property of the model.

Discriminative kernel convolution network for multi-label ophthalmic disease detection on imbalanced fundus image dataset.

It is feasible to recognize the presence and seriousness of eye disease by investigating the progressions in retinal biological structures. Fundus examination is a diagnostic procedure to examine the biological structure and anomalies present in the eye. Ophthalmic diseases like glaucoma, diabetic retinopathy, and cataracts are the main cause of visual impairment worldwide. Ocular Disease Intelligent Recognition (ODIR-5K) is a benchmark structured fundus image dataset utilized by researchers for multi-label multi-disease classification of fundus images. This work presents a Discriminative Kernel Convolution Network (DKCNet), which explores discriminative region-wise features without adding extra computational cost. DKCNet is composed of an attention block followed by a Squeeze-and-Excitation (SE) block. The attention block takes features from the backbone network and generates discriminative feature attention maps. The SE block takes the discriminative feature maps and improves channel interdependencies. Better performance of DKCNet is observed with InceptionResnet backbone network for multi-label classification of ODIR-5K fundus images with 96.08 AUC, 94.28 F1-score, and 0.81 kappa score. The proposed method splits the common target label for an eye pair based on the diagnostic keyword. Based on these labels, over-sampling and/or under-sampling are done to resolve the class imbalance. To check the bias of the proposed model towards training data, the model trained on the ODIR dataset is tested on three publicly available benchmark datasets. It is observed that the proposed DKCNet gives good performance on completely unseen fundus images also.

A deep CNN model for anomaly detection and localization in wireless capsule endoscopy images.

Wireless capsule endoscopy (WCE) is one of the most efficient methods for the examination of gastrointestinal tracts. Computer-aided intelligent diagnostic tools alleviate the challenges faced during manual inspection of long WCE videos. Several approaches have been proposed in the literature for the automatic detection and localization of anomalies in WCE images. Some of them focus on specific anomalies such as bleeding, polyp, lesion, etc. However, relatively fewer generic methods have been proposed to detect all those common anomalies simultaneously. In this paper, a deep convolutional neural network (CNN) based model 'WCENet' is proposed for anomaly detection and localization in WCE images. The model works in two phases. In the first phase, a simple and efficient attention-based CNN classifies an image into one of the four categories: polyp, vascular, inflammatory, or normal. If the image is classified in one of the abnormal categories, it is processed in the second phase for the anomaly localization. Fusion of Grad-CAM++ and a custom SegNet is used for anomalous region segmentation in the abnormal image. WCENet classifier attains accuracy and area under receiver operating characteristic of 98% and 99%. The WCENet segmentation model obtains a frequency weighted intersection over union of 81%, and an average dice score of 56% on the KID dataset. WCENet outperforms nine different state-of-the-art conventional machine learning and deep learning models on the KID dataset. The proposed model demonstrates potential for clinical applications.

Detection of abnormality in wireless capsule endoscopy images using fractal features.

One of the most recent non-invasive technologies to examine the gastrointestinal tract is wireless capsule endoscopy (WCE). As there are thousands of endoscopic images in an 8-15 h long video, an evaluator has to pay constant attention for a relatively long time (60-120 min). Therefore the possibility of the presence of pathological findings in a few images (displayed for evaluation for a few seconds only) brings a significant risk of missing the pathology with all negative consequences for the patient. Hence, manually reviewing a video to identify abnormal images is not only a tedious and time consuming task that overwhelms human attention but also is error prone. In this paper, a method is proposed for the automatic detection of abnormal WCE images. The differential box counting method is used for the extraction of fractal dimension (FD) of WCE images and the random forest based ensemble classifier is used for the identification of abnormal frames. The FD is a well-known technique for extraction of features related to texture, smoothness, and roughness. In this paper, FDs are extracted from pixel-blocks of WCE images and are fed to the classifier for identification of images with abnormalities. To determine a suitable pixel block size for FD feature extraction, various sizes of blocks are considered and are fed into six frequently used classifiers separately, and the block size of 7×7 giving the best performance is empirically determined. Further, the selection of the random forest ensemble classifier is also done using the same empirical study. Performance of the proposed method is evaluated on two datasets containing WCE frames. Results demonstrate that the proposed method outperforms some of the state-of-the-art methods with AUC of 85% and 99% on Dataset-I and Dataset-II respectively.

Functional grouping of similar genes using eigenanalysis on minimum spanning tree based neighborhood graph.

Gene expression data clustering is an important biological process in DNA microarray analysis. Although there have been many clustering algorithms for gene expression analysis, finding a suitable and effective clustering algorithm is always a challenging problem due to the heterogeneous nature of gene profiles. Minimum Spanning Tree (MST) based clustering algorithms have been successfully employed to detect clusters of varying shapes and sizes. This paper proposes a novel clustering algorithm using Eigenanalysis on Minimum Spanning Tree based neighborhood graph (E-MST). As MST of a set of points reflects the similarity of the points with their neighborhood, the proposed algorithm employs a similarity graph obtained from k(') rounds of MST (k(')-MST neighborhood graph). By studying the spectral properties of the similarity matrix obtained from k(')-MST graph, the proposed algorithm achieves improved clustering results. We demonstrate the efficacy of the proposed algorithm on 12 gene expression datasets. Experimental results show that the proposed algorithm performs better than the standard clustering algorithms.