Ringkasan Portfolio
A real-time interactive puzzle game built with Python, OpenCV, and MediaPipe / cvzone.
This application captures a photo of the user via webcam, slices it into a 3x3 puzzle grid, and allows the player to solve the puzzle using hand gesture tracking (pinch-to-grab mechanic) without touching the mouse or keyboard!
Deskripsi Detail
🧩 Gesture-Based Photo Puzzle Game (Computer Vision / AI)
Tech Stack: Python, OpenCV, MediaPipe, cvzone, NumPy
An interactive real-time puzzle game that turns a webcam snapshot into a 3x3 grid puzzle, played entirely using computer vision and hand gesture controls without touching a mouse or keyboard.
Key Features & Technical Highlights:
Real-time Hand Tracking: Integrated `cvzone` & MediaPipe for low-latency hand landmark detection and tracking.
Custom Gesture Recognition: Implemented mathematical distance detection between hand landmarks to calculate "pinch" gestures for grabbing, dragging, and dropping puzzle pieces.
Snap-to-Grid & Game Logic: Built custom spatial collision algorithms to handle dynamic puzzle piece placement, locking mechanisms, and victory state detection.
Clean UI State Handling: Managed frame rendering states to isolate raw photo capture mode from interactive tracking overlays.
GitHub Repository: gesture_puzzle_python
