Deep learning-based analysis of COVID-19 X-ray images: Incorporating clinical significance and assessing misinterpretation 

Journal : Digital Health  

In this study, (COV-X-net19) a convolutional neural network model is developed. Though there were numerous DL models for the classification of COVID-19, assessing the misinterpretation, explaining why some of the images are misclassified, and finding out the reason for this misclassification with statistical modeling was my motivation behind this work. 

Improving the Automated Diagnosis of Breast Cancer with Mesh Reconstruction of Ultrasound Images Incorporating 3D Mesh Features and a Graph Attention Network

Journal: Journal of Imaging Informatics in Medicine 

The main objective of this study is to develop a novel approach for breast tumor classification using 3D mesh reconstructed from 2D ultrasound images. By converting the region of interest into a 3D representation, we aim to enhance tumor analysis and improve classification accuracy. The proposed method, incorporating a Graph Attention Network (GAT), significantly outperforms traditional models, offering a more reliable and efficient diagnostic system for radiologists. 

Enhancing Autism Spectrum Disorder Classification with Lightweight Quantized CNNs and Federated Learning on ABIDE-1 Dataset

Journal: Mathematics, MDPI
This study aims to improve ASD classification using advanced deep learning techniques applied to neuroimaging data. We developed an automated system leveraging the ABIDE-1 dataset and a novel lightweight quantized one-dimensional (1D) Convolutional Neural Network (Q-CNN) model to analyze fMRI data. 

Malignancy pattern analysis of breast ultrasound images using clinical features and a graph convolutional network

Journal: Digital Health  

This study proposes an automated system for breast cancer classification using clinical markers such as tumor shape, margin, and surrounding tissue. A graph convolutional network (GCN) is developed to classify tumors into benign or malignant, achieving 98.73% accuracy. The GCN model's robustness is tested across three additional breast ultrasound datasets, yielding strong results. The study highlights the importance of using clinically relevant features in breast cancer diagnosis. 

Graph neural network-based breast cancer diagnosis using ultrasound images with optimized graph construction integrating the medically significant features

Journal : Journal of Cancer Research and Clinical Oncology 

This study develops an automated approach for classifying breast tumors into benign and malignant using ultrasound images. A novel Graph Neural Network (GNN) model is proposed, utilizing clinically significant features extracted from regions of interest (ROI). This approach aims to improve early diagnosis accuracy and assist radiologists in effectively identifying breast cancer patterns.  

An automated privacy-preserving self-supervised classification of COVID-19 from lung CT scan images minimizing the requirements of large data annotation 

Journal: Journal of Cancer Research and Clinical Oncology 

This study presents a novel privacy-preserving self-supervised (SSL) framework for COVID-19 classification from lung CT scans, utilizing federated learning (FL) enhanced with Paillier homomorphic encryption (PHE) to prevent third-party attacks during training. Training is done using the Bootstrap Your Own Latent (BYOL) contrastive learning SSL framework with a VGG19 encoder followed by attention CNN blocks (VGG19 + attention CNN).   

A Low Complexity Efficient Deep Learning Model for Automated Retinal Disease Diagnosis

Journal: Journal of Healthcare Informatics Research

This paper classifies human retinal OCT images into four classes using a lightweight optimized compact convolutional transformers (OCCT) model. Data imbalance in the Drusen and DME classes is addressed with GAN-based augmentation, creating a dataset of 130,649 enhanced images. The OCCT model, achieving 97.09% accuracy, outperforms ViT, Swin Transformer, and eight transfer learning models, even with reduced training images. The study highlights the OCCT model’s effectiveness and stability for retinal condition classification using deep learning