This can support gesture-based inputs or region-based classification where the user can point at objects.
Key features:-
- Use MediaPipe's hand tracking solution
- Overlay detected hand landmarks on the image or video
- (Optional) Map gestures to trigger predictions
Adding gesture-based input or using hands to segment garbage makes the model more interactive and could be useful in smart bin prototypes or assistive tech.
Kindly assign me this issue to contribute upon, under GSSoC 2025.