Analysis updated 2026-07-31 · repo last pushed 2021-02-05
Track multiple small targets across surveillance video footage automatically.
Analyze sports clips by following players or objects as they move throughout the video.
Monitor small moving objects in a video stream where detection-from-scratch every frame is too slow.
| jiang-cx/yolov4_siamfcpp_multitracking | 0xallam/posthog | 0xallam/search-engine | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | — | Python | C++ |
| Last pushed | 2021-02-05 | 2026-03-26 | 2023-08-23 |
| Maintenance | Dormant | Maintained | Dormant |
| Setup difficulty | hard | moderate | hard |
| Complexity | 4/5 | 3/5 | 3/5 |
| Audience | researcher | pm founder | developer |
Figures from each repo's GitHub metadata at analysis time.
README lacks setup instructions, you need familiarity with detection and tracking frameworks plus compatible model weights.
This project is a tool for automatically following multiple small objects as they move around in a video. It combines two techniques: one to detect the objects, and another to keep tracking them efficiently over time. The approach pairs YOLOv4, which scans each frame to locate tiny objects, with SiamFC++, which handles the tracking step in a way designed to be faster than starting detection from scratch on every frame. The "online" part means it processes frames as they come in, without looking ahead at future video. Someone might use this for scenarios like monitoring small targets in surveillance footage or analyzing sports clips. The README doesn't go into detail about setup, supported video formats, or performance benchmarks, so practical guidance is limited. You'd need some familiarity with object detection and tracking frameworks to get it running.
A video tool that detects and tracks multiple small moving objects by combining YOLOv4 detection with SiamFC++ tracking. It processes frames one at a time without looking ahead.
Dormant — no commits in 2+ years (last push 2021-02-05).
No license information is provided in the README, so you should contact the author before using this code.
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly researcher.
This repo across BitVibe Labs
Verify against the repo before relying on details.