Brain tumors are a potentially deadly form of cancer that result in thousands of new cases every year. MRI scans are conducted on the human brain in order to identify, visualize, and follow up on these tumors (if present). Although MRI scans are a very effective method in identifiy and classifying brain tumors, mistakes are still possible. Tumors may be misinterpreted or not detected at all, which can allow them to grow into more intense cancer cases. Our code will scan an MRI brain scan inputted by the user, and using a combination of denoising, filtering, and image thresholding, will identify the tumor. If no tumor is present, it will return a message stating so. If a tumor is present, it will state so and output a new image with the tumor outlined.
The code will ask the user to input the image name, then read the image and save it to a variable called 'img'. The code will then utilize the function title 'project_function' to denoise the image, without taking out any other important parts of the image. Then the image will be resized and thresholded it to isolate areas that may be tumors from the rest of the scan. The code then creates a box over the area that may be a tumor, and allows this individual area to be scanned in order to determine if it is a tumor, based off its shape and area. The code then sets an area of 'high density' as that above 0.5. We determined this would be the appropriate value after testing many of the images in the folders, and seeing what value would be classified as 'high density' for a brain tumor. Due to this, our code outlines the areas of highest density, so that other areas of the scan that may only appear to be a tumor are not highlighted. Then we determine if the are is considered high enough density by comparing it to the value we set (140). If a tumor is present, it is then outlined. At the end, the code either states that no tumor is present, or that one is present, in which case it will outline it in red and display it alongside the original image.