How althafshaik15/realvision works
RealVision is a webcam-based object identifier. The user starts the app and can pause, mute speech, request a scan, or switch detection backends; captured frames are analyzed, results are drawn over the video, and selected labels can be spoken aloud. The source shows local YOLO and optional Gemini recognition, but no README or sampled source documents further product features.
Main entry points: main.py App loop · detector.py Detector · speaker.py Speech selection · config.py Configuration
Components 12
App workflow
- App loop
main.py - Configuration
config.py
Object recognition
- Detector
detector.py - YOLO backend
detector.py - Gemini backend
detector.py - YOLO weights
yolov8n.pt
Result presentation
- Video display
- Speech selection
speaker.py
Other components
- User
- Webcam
- Gemini API
- System speech
Connections 13
- User → App loop: controls
- Webcam → App loop: provides frames
- App loop → Configuration: reads settings
- App loop → Detector: submits frames
- Detector → YOLO backend: dispatches
- Detector → Gemini backend: dispatches
- YOLO backend → YOLO weights: loads weights
- Gemini backend → Gemini API: sends image
- Detector → App loop: returns detections
- App loop → Video display: shows video
- App loop → Speech selection: submits detections
- Speech selection → Configuration: reads speech settings
- Speech selection → System speech: speaks labels