How skyao2002/real-time-object-detection works
This is a browser-based hand-pose demo: after loading the model, the app reads webcam frames, runs detection repeatedly, and draws labeled bounding boxes on a canvas. The sampled bundle supports this workflow; the model and webcam components are bundled dependencies, so their internal behavior is not shown.
Main entry points: main.6da63bb4.chunk.js Detection app · 2.ba583733.chunk.js Handpose model
Components 7
Video capture
- Webcam component
static/js/2.ba583733.chunk.js - Browser camera
Detection and display
- Detection app
static/js/main.6da63bb4.chunk.js - Handpose model
static/js/2.ba583733.chunk.js - Box renderer
static/js/main.6da63bb4.chunk.js - Overlay canvas
static/js/main.6da63bb4.chunk.js
Other components
- User
Connections 8
- User → Detection app: opens demo
- Detection app → Webcam component: renders
- Webcam component → Browser camera: captures video
- Detection app → Handpose model: loads model
- Detection app → Handpose model: submits frames
- Handpose model → Detection app: returns detections
- Detection app → Box renderer: draws detections
- Box renderer → Overlay canvas: paints boxes