Automatic Detection of Brain Tumor from CT and MRI Images using Wireframe model and 3D Alex-Net

Automatic detection of brain tumors from CT and MRI images is always an effortful task because of the complexity and heterogeneous images. Many neural networks architecture (NN) have recently been developed for segmentation and classification tasks and have proved quite successful. Studies that have taken into account the sizes of items have been rare; as a result, the majority of them show poor detection performance for tiny objects. This has the potential to have a significant influence on illness identification. Recently, the 3D neural network became popular because it can work with a large labeled dataset. We proposed a 3D Alex-Net-based architecture that can classify the different types of a brain tumors at an early stage. First, the image contour is identified and given to the classifier for class-wise identification. We tested our proposed approach on RSNA- MICCAI brain tumors and found that the proposed method delivers the highest accuracy, and the results provide a clear advantage for the classification of a brain tumor in medical images.