Processing of MRI images for detection of brain tumors

Abstract

We used the MRI technique to achieve a dataset of ten images: five healthy brain images and five images of brains with tumors. To highlight tumor development, the images were preprocessed using OpenCV, converted into grayscale, then underwent contrast changes, resizing, and Gaussian blur. The feature extraction stage involved binarizing images in well-defined values and extracting the tumor contour using the "find_contours" and "grab_contours" functions. The segmentation stage added the contour to a copy of the resized image, highlighted it with endpoints, and cropped the image to the edges of the contour. The results showed the changes in the tumor area after contrast and well-defined tumor tissues after feature extraction and segmentation.