/VisualRFS

This is the official Python and C++ implementation repository for a paper entitled "Visual multi-object tracking with re-identification and occlusion handling using labeled random finite sets", Pattern Recognition (https://arxiv.org/pdf/2407.08872).

Primary LanguagePythonMIT LicenseMIT

Visual RFS

This is the official Python and C++ implementation repository for a paper entitled "Visual Multi-Object Tracking with Re-Identification and Occlusion Handling using Labeled Random Finite Sets", Pattern Recognition (arXiv 2405.18606), (MOTChallenge).

Quick Overview

  • Exploiting object feature improves tracking performance of labeled random finite set filters (GLMB and LMB filters) by improving data association.
  • Using object deep feature is effective for re-ID tasks.
  • Fuzzy occlusion handling model improves tracking performance.

News

  • Our implementation for 3D multi-camera multi-object tracking with re-identification is released at 3D-Visual-MOT
  • Our GLMB/LMB Python package implementations adapted from Matlab, are released at labeledRFS

Docker image

Docker image to run VisualRFS codes can be found in VisualRFS Docker Hub.

Usage

  1. Set Up Python Environment

    • Create a conda Python environment and activate it: 
      conda create --name virtualenv python==3.7.16
conda activate virtualenv
    • Clone this repository recursively to have fuzzylite, and pybind11 
      git clone --recursive https://github.com/linh-gist/visualrfs.git

  2. Install Packages

    • C++ Packages (Make sure Eigen 3.4.0 is installed): Navigate to the src/cpputils folder and run python setup.py build develop
    • Python Packages: pip install -r requirements.txt

  3. Configure Experiment Options
    Change the following options when initializing GLMB or LMB tracker

    • C++ GLMB(int width, int height, bool useFeat = true, bool useFuzzyPD = false)
    • C++ LMB(int width, int height, bool useFeat = true, bool useFuzzyPD = false)
    • Python LMB: def __init__(self, use_feat=True)
    • Python GLMB: def __init__(self, width, height, use_feat=True)

  4. Prepare Data

    • Datasets:
      • MOTChallenge datasets (MOT16, MOT17, MOT20)
      • You can run on your own dataset with your preference object detector
    • Folder structure: 
      |-- detection
|   |-- detector_fairmot256
|   |   |-- MOT16-01.npz
|   |   |-- MOT16-02.npz
|   |   |   ...
|   |   |-- MOT16-14.npz
|   |-- detector_fairmot256
|   |   |-- MOT20-01.npz
|   |   |-- ...
|   |   |-- MOT20-08.npz
|-- src
|   |-- cpputils
|   |-- joint_glmb
|   |-- joint_lmb
|   |-- tracking_utils
|-- requirements.txt
|-- README.md
    • OSPA2 is re-implemented in Python and following this paper, an example code is given in ospa2.py. 
      @article{rezatofighi2020trustworthy,
  title={How trustworthy are the existing performance evaluations for basic vision tasks?},
  author={Tran Thien Dat Nguyen and Hamid Rezatofighi and Ba-Ngu Vo and Ba-Tuong Vo and Silvio Savarese and Ian Reid},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year={2022}
}

  5. Run the Tracking Demo

    • Navigate to the joint_glmb or joint_lmb and run python run_joint_glmb.py OR python run_joint_lmb.py.

Contact

Linh Ma (linh.mavan@gm.gist.ac.kr), Machine Learning & Vision Laboratory, GIST, South Korea

Citation

If you find this project useful in your research, please consider citing by:

@article{van2024visual,
      title={Visual Multi-Object Tracking with Re-Identification and Occlusion Handling using Labeled Random Finite Sets}, 
      author={Linh~Van~Ma and Tran~Thien~Dat~Nguyen and Changbeom~Shim and Du~Yong~Kim and Namkoo~Ha and Moongu~Jeon},
      journal={Pattern Recognition},
      volume = {156},
      year={2024},
      publisher={Elsevier}
}