https://github.com/fcakyon/mmdetection-object-tracker

A lightweight script for performing Kalman filter based object tracking using MMDetection models.
https://github.com/fcakyon/mmdetection-object-tracker

colab kalman-tracking mmdet mmdetection norfair notebook object-detection object-tracking python pytorch realtime-tracking torch tracking varifocal-loss varifocalnet vfnet

Last synced: about 2 months ago
JSON representation

A lightweight script for performing Kalman filter based object tracking using MMDetection models.

• Host: GitHub
• URL: https://github.com/fcakyon/mmdetection-object-tracker
• Owner: fcakyon
• License: mit
• Created: 2021-06-23T18:37:29.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2021-06-26T09:35:33.000Z (over 3 years ago)
• Last Synced: 2024-12-26T20:51:33.613Z (about 2 months ago)
• Topics: colab, kalman-tracking, mmdet, mmdetection, norfair, notebook, object-detection, object-tracking, python, pytorch, realtime-tracking, torch, tracking, varifocal-loss, varifocalnet, vfnet
• Language: Python
• Homepage:
• Size: 27.3 KB
• Stars: 23
• Watchers: 4
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        




  MMdetection + Tracker





  Simplest possible example of tracking. Based on MMDetection Detector.









## 
Instructions


1. Install Norfair with `pip install norfair[video]`.

2. Install MMDetection with `pip install torch mmcv-full mmdet`.

3. Run `python demo.py `.

4. Bonus: Use additional arguments `--model_path`, `--config_path`,`--img_scale`, `--conf_thres`, `--max_track_distance`, `--track_points` as you wish.

## 
Explanation


This example tracks objects using a single or double point per detection: the centroid or the two corners of the bounding boxes around objects returned by MMDetection Detector.

### VFNet (MMDetection) tracking demo: