Gradient-weighted Class Activation Mapping (Grad-CAM) is a class-discriminative localization technique that can generate visual explanations from any CNN-based network without any architectural requirement or re-training. This approach uses the gradient flowing into the final convolutional layer to produce a localization map highlighting the important regions in the image for predicting the specific target.
The aim of the project it to apply Grad-CAM technique to a CNN trained on a brain tumor classification task and evaluate the visual explanation.
The model used for the classification is ResNet50, which has been fine tuned by using the "Brain MRI images for brain tumor detection" dataset.
The results of the experiment were highly satisfactory, demonstrating a remarkable ability to accurately localize the tumor's exact region as we can see in figure below.
