Stack
Python
OpenCV
◆MediaPipe
◆Input automation
← Back to portfolio
Case study
Camera Mouse
- Computer vision
- Accessibility
- Real-time control
Hands-free controller that maps webcam input to cursor motion and actions with OpenCV and MediaPipe—built to explore low-latency gesture pipelines on everyday hardware.
Preview
Demo clips and calibration notes ship with the repo—local run is the best preview.
Impact
Pointer control without a desk-bound device is useful for accessibility experiments, rapid prototyping, and proof-of-concept ergonomic setups. The project focuses on stable landmark tracking and sane defaults so the loop stays responsive.
Architecture
Capture: webcam frames through OpenCV with tuned resolution and FPS targets for latency.
Understanding: MediaPipe hands/ pose pipelines to derive screen mapping from gestures.
Control: OS-level mouse events with smoothing and dead zones to reduce jitter.
Shipped
- Gesture-to-cursor pipeline with configurable sensitivity
- Modular capture and inference stages
- Documentation for dependencies and camera permissions
- Foundation for additional gestures or accessibility profiles