The paper presents a deep learning model for the detection of anterior cruciate ligament (ACL) tears in knee MRI scans, using transfer learning with Convolutional Neural Networks (CNNs). The model leverages pretrained networks and fine-tunes them for ACL tear classification. To enhance the interpretability of the model, Class Activation Maps (CAMs) are employed to visualize the discriminative regions in the MRI scans that contribute to the predictions, offering insights into the model’s decision-making process. The proposed system shows promising results, with an accuracy of 87.6%, demonstrating both high performance and the potential for clinical deployment.