Lumbar Spine Disease Detection

Abstract: Detection of lumbar diseases using deep learning is the first step toward the application of artificial intelligence technology in medical diagnostics. This project provides the INVGG-based diagnostic system for spinal health problems. Its primary aim is to develop an infrastructure that is fully self-sufficient, inexpensive, and efficiently classifies lumbar conditions such as herniated discs and spinal stenosis. The model is augmented with features such as normalization preprocessing, learning enhancement augmentation, and visual understanding explanation via Grad-CAM with minimal human monitoring. The use of deep learning allows the remote diagnosis and monitoring of spinal disorders. The system includes a web interface for immediate prediction of diseases, displaying heatmaps for diagnostics which helps radiologists and medical practitioners to provide timely, accurate, and efficient diagnoses. This system allows uploading images, displaying results alongside generated reports thus increasing operational efficiency. This spinal diagnostic system presents an adaptive solution to strengthen the spine healthcare system while improving the initiatives for clinical and research endeavors focused on lumbar diseases.