Hariharan Sureshkumar

Hariharan Sureshkumar

Computer Vision Researcher, Vision-only Robotic Perception, 3D Vision

Classical Computer Vision Interactive Web Interface

Project Overview

A comprehensive repository and interactive web application containing implementations of classical computer vision algorithms, research papers, and a Gradio-based web interface for testing these algorithms on custom images. This project serves as both a learning resource and a practical toolkit for classical computer vision.

Technologies Used

  • Python (OpenCV, NumPy, PIL/Pillow)
  • Gradio for interactive web interface
  • Scikit-image for additional image processing
  • Matplotlib for visualization
  • Computer Vision Algorithms: SIFT, SURF, ORB, Canny, Harris corner detection

Key Features

  • Interactive Web Interface: Upload images and test various CV algorithms through an intuitive web interface
  • Comprehensive Algorithm Library: Implementation of fundamental CV algorithms including:
    • Image processing (Gaussian blur, median filter, bilateral filter, sharpening)
    • Edge detection (Sobel, Canny, Harris corner detection, Hough transforms)
    • Feature detection (SIFT, SURF, ORB, BRIEF)
    • Segmentation (Thresholding, watershed, SLIC superpixels)
    • Object detection (HOG, Haar cascades, template matching)
  • Educational Content: Organized tutorials covering all major CV topics
  • Research Papers Collection: Curated collection of seminal papers in computer vision

Algorithm Categories Covered

  1. Image Fundamentals: Pixel intensity, color spaces, image formats, histograms
  2. Image Preprocessing: Noise filtering, sharpening, histogram equalization, gamma correction
  3. Edge and Feature Detection: Gradient operators, Canny edge detector, corner detection
  4. Feature Descriptors and Matching: SIFT, SURF, ORB, descriptor matching algorithms
  5. Image Segmentation: Thresholding, morphological operations, watershed, superpixel algorithms
  6. Object Detection and Recognition: Template matching, Haar cascades, HOG
  7. Motion and Optical Flow: Frame differencing, background subtraction, optical flow
  8. Camera Geometry and Calibration: Pinhole camera model, camera calibration
  9. Stereo Vision and Depth Estimation: Stereo matching, disparity computation
  10. Structure from Motion and 3D Vision: Feature tracking, bundle adjustment, point clouds
  11. Image Registration and Stitching: Image alignment, panorama stitching

Research Contributions

  • Comprehensive implementation of classical computer vision algorithms
  • Interactive learning platform for computer vision education
  • Modular architecture for easy algorithm testing and comparison
  • Real-time visualization of algorithm effects on custom images

Live Demo

Open Interactive Demo

Click the button above to open the interactive Classical Computer Vision demo in a new tab

GitHub

Hugging Face Space