Ten ways to play with your webcam — because I like understanding how computers see the world, and I like turning that into things people actually want to run.
This is my personal playground of camera + computer-vision experiments. Each project here is a real, runnable thing: point your webcam at yourself and play — make the world glitch when you open your mouth, conduct a virtual orchestra with your arms, or turn a Rubik's Cube in front of your camera and watch it get solved.
Every project has two sides:
- Play Mode — clean and fun, for anyone with a webcam.
- Experiment Mode — press a key and see what the computer actually sees: landmarks, confidence, FPS, latency. For the curious (and great for demos).
| Project | What it does | Status |
|---|---|---|
| Vision Filters | Effects driven by your face — open your mouth and the world glitches | 🚧 building |
| MemeVision | Turn yourself into a meme: expressions, gestures, sounds, captions | 🗺️ on the map |
| Project | What it does | Status |
|---|---|---|
| GestureOS | Control your computer with gestures — mouse, media, windows | 🗺️ on the map |
| Vision Games | The camera becomes a game controller | 🗺️ on the map |
| Project | What it does | Status |
|---|---|---|
| Motion Studio | Pose estimation makes movement fun — reps, dance, boxing | 🗺️ on the map |
| Vision Music | Your body becomes a musical instrument — the AI Conductor | 🗺️ on the map |
| Project | What it does | Status |
|---|---|---|
| Creative Studio | Create art with the camera — slit-scan, particles, fluid | 🗺️ on the map |
| Vision Challenges | Point the camera at a puzzle and solve it — Rubik's, chess, Sudoku | 🗺️ on the map |
| Vision Lab | The camera as a measuring tool — OCR, depth, scanning | 🗺️ on the map |
| Project | What it does | Status |
|---|---|---|
| AI Vision | Give the camera a brain — private, local AI that sees you | 🗺️ on the map |
flowchart LR
CAM[📷 Your webcam] --> FACE[😀 Face<br/>expressions & pose]
CAM --> HANDS[🖐️ Hands<br/>gestures]
CAM --> BODY[🧍 Body<br/>movement]
CAM --> SCENE[🌍 Scene<br/>objects & depth]
FACE --> FILTERS[Vision Filters]
FACE --> MEMES[MemeVision]
HANDS --> GESTURES[GestureOS]
HANDS --> GAMES[Vision Games]
BODY --> MOTION[Motion Studio]
BODY --> MUSIC[Vision Music]
SCENE --> ART[Creative Studio]
SCENE --> PUZZLES[Vision Challenges]
SCENE --> LAB[Vision Lab]
CAM --> AI[AI Vision]
CORE[🧱 shared core<br/>camera + tracking + smoothing + events] -.-> FILTERS
CORE -.-> GAMES
CORE -.-> MOTION
CORE -.-> MUSIC
Everything shares one small, honest core — the parts that would be painful to rewrite (camera capture, tracking, smoothing, event handling) live once, and each project builds its own personality on top of it.
- The boring engineering is the point. Interactions feel stable and intentional, not twitchy and demo-like: One-Euro smoothing, hysteresis, confidence gates, cooldowns, state machines. Most of the code is making signals behave, not detecting them.
- Honest measurements. Where I measure (depth, distances, counting), I publish the error bars. Understanding what computer vision can't do is part of the craft.
- Python and TypeScript, used where they fit. Python for the OS-level and heavy-CV work, the browser for the things you should be able to try with zero install.
- Every project runs locally. Your camera feed stays on your machine.
- Now: Vision Filters — the landmark-driven Shader Lab is live; effects driven by your face (glitch on mouth-open, kaleidoscope on eyebrow-raise, and more coming).
- Next: Creative Studio, MemeVision, then AI Vision.
- Then: games, music, puzzles, and the rest — one fun, polished thing at a time.
The full roadmap and idea bank live in my private planning. This page grows with the work.
Made with curiosity by Muneer Alam — a BITS Pilani student who likes understanding how computers see the world.
