Computer Vision Suite
Three real-time OpenCV tools: a gesture paint app, a document scanner, and a live license-plate detector.
A trio of computer-vision applications built on a shared C++/OpenCV foundation, each targeting a different classic vision problem: color tracking, document rectification, and object detection.
- Virtual Paint: real-time color detection + contour tracking turns physical objects into digital markers.
- Document Scanner: a Canny → perspective-transform pipeline rectifies skewed images into clean scans.
- License-Plate Detector: Haar cascades localize and track plates in live video feeds.
01Virtual Paint
Real-time color segmentation isolates a colored marker in the camera feed; contour tracking follows its centroid frame-to-frame, letting you draw on a digital canvas by moving a physical object through the air.
02Document Scanner
An image-processing pipeline detects page edges with Canny, finds the document contour, and applies a perspective transform to warp a skewed photo into a flat, top-down scan.
03License-Plate Detector
A Haar-cascade classifier localizes plates in a live video stream and tracks them across frames, the same detection approach used in classic real-time object-detection systems.
Want to talk through how I built this?