Raspberry Pi Powered Deep Learning-based Chest X-ray Analysis for Medical Diagnosis

Despite the impressive performance, traditional medical imaging analysis equipment is often costly and cumbersome, limiting accessibility, especially in remote or resource-constrained areas. This work proposes harnessing the capabilities of deep learning models, particularly Convolutional Neural Networks (CNNs). This study explores the feasibility of employing CNN models, particularly ResNet family, which have demonstrated exceptional performance in various deep learning applications. In this work, CNN is applied for the detection of lung diseases by considering chest X-rays and automatically classifying them into four categories: COVID-19, pneumonia, lung opacity, and normal. Furthermore, the research investigates the implementation of the trained CNN model on a Raspberry Pi, a low-cost, single-board computer with decent computational power. By deploying the model on such an accessible platform, the aim is to extend the benefits of deep learning-based chest X-ray analysis to resource-limited environments. The proposed model was able to achieve 97.98% classification accuracy and demonstrated significant potential in classifying chest X-ray images. The integration of efficient deep learning architectures contributes to the development of robust diagnostic capabilities, potentially revolutionizing healthcare accessibility in underserved regions.