asigatchov/volleyball-tracking

Fast Volleyball Ball Tracking Demo

This demo showcases real-time detection and tracking of a volleyball at 80 FPS on CPU using an optimized ONNX model. Upload a volleyball video to see the ball coordinates in CSV format and optionally visualize the tracking output.

How to Use

1. 1.Upload a volleyball video via the interface.
2. 2.Adjust the track length (default: 10 frames).
3. 3.Check the "Visualize Output" box to see the ball trajectory.
4. 4.Click "Submit" to start inference.

About the Project

• —Repository: asigatchov/fast-volleyball-tracking-inference
• —Model: VballNetFastV1 (ONNX)
• —Performance: 200 FPS on Intel Core i5-10400F

Example Output

[image]

MIT License