Please provide python code of this question

a. [50 points] Implement the Harris keypoint detector as discussed in the lectures. You do not need to implement a descriptor. Apply it to the two images provided and threshold the corner- ness response so that about a few hundred keypoints are returned. Submit a result showing the (top-scoring) detected Harris keypoints superimposed on the original image (e.g., youcan place a cross or a circle at the location of the keypoint; do this on the grayscale version to facilitate visualization). Report the threshold that you choose. Describe which objects or regions in the image seem to generate large numbers of Harris keypoints, and why. You may not use the built-in library function for Harris detection, but must implement it on your own. b. [25 points] Rotate the image in increments of 15, from 0 to 360. You are allowed use of library function(s) for rotation. For each rotated image, compute Harris keypoints using the same settings that you chose in part (a). Then, compute the repeatability using the following procedure: . Set number of feature matches M to be 0 . For each Harris keypoint at [x y] in the original image: o Predict the position Xt = [X y] where the keypoint should ideally appear in the rotated image (by applying a rotation transformation on [x y]). o Search for a nearby Harris keypointwhich is detected in the rotated image with coordinates xo = [xo, yo] satisfying Ilxo - Xill s T, where T is some suitably chosen threshold. o If such an xo = [xo, yo] is found, increment number of feature matches M by 1. This matching point xo in the rotated image should now not be considered for matching in subsequent iterations. . Compute repeatability as M/N, where M is the number of feature matches and / is the number of Harris keypoints in the original image. Plot repeatability against rotation angle and comment on the Harris keypoint detector's robustness against in-plane rotation. c. [25 points] Conduct an experiment analogous to part (b), but instead of rotating the image, resize the image by the scaling factors NO, N', N2, ... N8, where m = 1.2. Compute Harris keypoints for each resized image. You are allowed to use built-in library functions to perform this scaling, preferably with bicubic interpolation. By comparing Harris keypoints of the original and resized images, compute and plot repeatability against scaling factor. Comment on the Harris keypoint detector's robustness against scale changes