Revealing Advanced Brain Tumour Detection: An In-Depth Study Leveraging Grad CAM Interpretability

Adult primary brain tumours are a deadly illness that can strike anyplace. With image processing techniques, the science of computer vision, and artificial intelligence (AI) in particular, has made great strides towards the automatic identification and detection of brain tumours. This study evaluated the ability of many deep learning models, such as VGG16, Inception V3, Mobile Net, ResNet 50 v2, and EfficientNet B1, to differentiate between brains that were tumour-free and those that were tumour-bearing using magnetic resonance imaging contrasts. As part of a strategy for explainable AI, techniques like Grad CAM are utilised for brain tumour localisation in order to further enhance the models’ interpretability. The findings imply that the application of explainable AI methods can improve deep learning models’ interpretability and could be essential for assessing how well they can identify brain tumours.