COMP425 Implementations

Overview

• Subject: COMP425/COMP6341 Computer Vision.
• Session: Winter 2024.
• Institution: Université Concordia.

Key Topics

• 2D Image Filtering.
• Edge and Corner Detection: corner response calculation, thresholding, and non-maximum suppression.
• Image Subsampling: Naive and anti-aliasing downsampling techniques.
• Hough Transform.
• Homography and RANSAC.
• Filter banks.
• K-Means Clustering for feature extraction.
• Bag-of-Words Representation.

Repository Contents

• HW1: Implementation of a 2D filtering, edge and corner detection and image sumsampling functions for images.

• HW2: First, implementation of Hough transform to find straight lanes in a road. Secondly, creating a pipeline for estimating the homography, and using the estimated homography in an AR (augmented reality) application.

• Visual Recognition System: Implementation of a simple texture recognition system using the bag-of-words (BoW) approach.