A LeakyReLU based Effective Brain MRI Segmentation using U-NET

Brain Tumor identification has been regarded as a critical topic. Meanwhile, it is complicated to spot the tumor in MRI images manually from a large amount of MRI images generated is difficult and time-consuming due to unpredictable shapes and sizes of the tumor. Image Segmentation techniques make a massive impact here and help in obtaining more significant results by dividing the image into segments for prior identification of tumors. U-Net with LeakyReLu can be used for faster and precise segmentation of medical images. Thresholding is applied to identify the ROI of the tumor for better identification of the abnormality of the tumor. Identifying the tumor region from the segmented MRI image is lesser time-consuming. Therefore, our model developed using neural networks can help the doctors in precisely identifying the tumor region from the segmented images and thereby assisting them to help the patients.