Lumbar Spine Degeneration Detection Using MRI Images

This study presents a deep learning-based system for classifying lumbar spine degeneration using MRI images. A hybrid Convolutional Neural Network (CNN) and Multi-Layer Perceptron (MLP) model was developed, integrating image-based features and coordinate data to detect the severity of Left Subarticular Stenosis. Transfer Learning techniques, including DenseNet, ResNet, and VGG16, were explored for performance enhancement. The model was evaluated using multiple metrics such as accuracy, precision, recall, AUC-ROC, and F1-score. Results indicate that the custom CNN-MLP hybrid model offers improved diagnostic accuracy compared to traditional methods. The findings emphasize the growing role of deep learning in automated medical image analysis.