CSE 527 Computer Vision - Fall 2017 - Stony Brook University
All implementations are written in Python 2.7 using OpenCV.
Contains an implementation of kernel deconvolution in the frequency domain as well as Laplacian Pyramid blending of two images.
Uses RANSAC to fit homography and affine transformations for image matching and panoramic stitching.
Implements a human face detector and tracker which detects faces using Viola-Jones. Tracking is accomplished using Camshift, Kalman filter, and particle filtering.
Uses SLIC to segment an image using superpixels, also includes an interactive foreground/background segmentation implementation which will segment based on user's input markings.
Reconstructs a sparse 3d point cloud of an image processed via structured light.