A comprehensive collection of Computer Vision implementations ranging from Classical Image Processing to Modern Deep Learning applications. This repository documents my exploration of how machines perceive, process, and interpret visual data.
- Description: Implements Gunnar Farneback’s algorithm to track pixel-level motion vectors in real-time via webcam.
- Tech: Python, OpenCV, NumPy.
- Highlights: Features dual-mode visualization using vector fields and HSV color mapping to represent motion direction and magnitude.
- Description: A library of fundamental CV techniques used for image analysis without heavy neural networks.
- Modules: * Edge Detection: Canny, Sobel, and Laplacian operators.
- Feature Matching: SIFT/ORB descriptors.
- Contour Analysis: Shape detection and object counting.
- Description: Leveraging MediaPipe and YOLO for tracking human movement and workspace occupancy.
- Hardware Integration: Includes Arduino-based LED triggers based on visual detection events.
- Languages: C++, Python
- Libraries: OpenCV, MediaPipe, NumPy, Matplotlib
- Frameworks: YOLO (You Only Look Once), TensorFlow/Keras
- Hardware: Arduino, USB Webcams, (Planned: ROV Thrusters for Underwater CV)