My Projects

Built a Haar Cascade–based face recognition system with 95% accuracy and a 50ms response time on images and live video. Optimized image preprocessing and tuning parameters to trim latency by 30%, sustaining performance under changing lighting conditions. Enhanced user flow with bounding boxes, highlighting detected faces in real time for instant feedback.

Engineered a computer vision–driven music control system, achieving 95% gesture accuracy and sub-200ms response times. Enabled real-time commands (pause/play) and cut track-switch latency from 1.2s to 850ms with open/closed-hand gestures. Facilitated intuitive track switching using thumb gestures for left and right navigation. Integrated OpenCV, Mediapipe, and Osascript to unify computer vision, hand tracking, and macOS scripting, creating a responsive user experience.