Enhanced Conjunctivitis Detection: Leveraging ResNet and Spatial Attention Mechanisms for Superior Accuracy

In recent years, leveraging deep learning techniques for medical image analysis has demonstrated considerable advancements. This study introduces an innovative method for detecting conjunctivitis by integrating a convolutional neural network (CNN) with a Spatial Attention Mechanism, specifically utilizing the ResNet architecture. Our approach aims to enhance diagnostic accuracy by focusing on critical features in ocular images. We preprocess the ocular images, apply normalization, and extract features using ResNet, followed by the application of a Spatial Attention Mechanism to highlight areas indicative of conjunctivitis. The dataset comprises 358 images, with 181 categorized as healthy and 177 showing signs of conjunctivitis. Impressively, the model obtained 92.5% accuracy, 90.7% sensitivity, 94.3% specificity, and an F1 Score of 0.92. With a low rate of false positives and a high specificity demonstrating a good capacity to identify afflicted cases, our results highlight the efficacy of our technique in accurately detecting conjunctivitis. The combination of ResNet with Spatial Attention Mechanism represents a significant advancement in automated eye disease diagnosis, offering a robust and precise tool for clinical applications.