EDIT MODE - Click a player to select
Detection Editor
Click a player on the video or in the list below to edit. Toggle edit mode to drag/resize boxes.
Track T1
Detection History
Player identifications will appear here
Settings
Model: basketball-player-detection (RF-DETR) — 10 classes incl. player, number, ball, rim
Tracker Settings (ByteTrack)
Live Detection
0
White Team
0
Blue Team
0
Roster Editor
00:00
Duration
0
Frames
0
IDs Confirmed
--
FPS
About this tool: This demonstrates the computer vision pipeline described in
"How to Detect, Track, and Identify Basketball Players with Computer Vision"
by Piotr Skalski (Roboflow, Sep 2025).
The original pipeline uses RF-DETR, SAM2, SigLIP, ResNet, and SmolVLM2.
This web adaptation uses Roboflow hosted inference and Moondream VLM.
Cite: Piotr Skalski. (Sep 30, 2025). How to Detect, Track, and Identify Basketball Players with Computer Vision. Roboflow Blog.
Cite: Piotr Skalski. (Sep 30, 2025). How to Detect, Track, and Identify Basketball Players with Computer Vision. Roboflow Blog.