/Computer-Vision

CSE 527 Computer Vision - Fall 2017 - Stony Brook University

Primary LanguagePython

Computer-Vision

CSE 527 Computer Vision - Fall 2017 - Stony Brook University

All implementations are written in Python 2.7 using OpenCV.

Gaussian Pyramids & Blending

Contains an implementation of kernel deconvolution in the frequency domain as well as Laplacian Pyramid blending of two images.

Image Stitching & Panoramas

Uses RANSAC to fit homography and affine transformations for image matching and panoramic stitching.

Tracking & Detection

Implements a human face detector and tracker which detects faces using Viola-Jones. Tracking is accomplished using Camshift, Kalman filter, and particle filtering.

Segmentation

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.

Structured Light 3d Reconstruction

Reconstructs a sparse 3d point cloud of an image processed via structured light.