Skip to content

SiamBOMB: Siamese network using Background information for real-time Online Multi-species home-cage animal Behavioral analysis. (Updating)

License

Notifications You must be signed in to change notification settings

YuanJingbin/SiamBOMB

Repository files navigation

SiamBOMB

License RepoSize CodeSize Release Commit

This repo is the second preview version of SiamBOMB, which is updating in September 2021.
Copyright (c) 2021 Institute of Automation, Chinese Academy of Sciences. All rights reserved.

Introduction

Our Paper (IJCAI 2020 Demo Track): 10.24963/ijcai.2020/776

  1. TODO

Setup

1. Configure environments

  • Linux (Ubuntu 18.04) or Windows (10).
  • GPU (at least have 4 GB memory).
  • CUDA 10.1, 10.2, 11.1, etc. (with cuDNN).
  • Anaconda 4.8+ (or virtualenv etc.) and Python 3.6+.
  • C++ build tools (g++ in Linux or 2015+ in Windows).
  • Download .zip or git clone our code.

2. Install dependencies

# create anaconda env
conda create -n SiamBOMB python=3.7
conda activate SiamBOMB

# install the pytorch
conda install pytorch torchvision cudatoolkit=11.1 -c pytorch -c conda-forge

# install the pyqt5
pip install pyqt5-tools

# install other packages
pip install numpy opencv-python pyyaml yacs tqdm colorama matplotlib cython tensorboardX imutils pandas tb-nightly visdom scikit-image
tikzplotlib spatial-correlation-sampler jpeg4py
sudo apt-get install ninja-build libturbojpeg

3. Equip models

  • SiamMask_E pretrained model: Google Drive, Baidu Pan (jffj)
    into pysot/experiments/siammaske_r50_l3/model.pth
  • KYS pretrained model: Google Drive
    into pytracking/networks/kys.pth
  • LWL pretrained model: Google Drive
    into pytracking/networks/lwl_boxinit.pth
  • KeepTrack pretrained model: Google Drive
    into pytracking/networks/keep_track.pth
    and base model: Google Drive
    into pytracking/networks/super_dimp_simple.pth

Citation

@inproceedings{SiamBOMB,
  title     = {SiamBOMB: A Real-time AI-based System for Home-cage Animal Tracking, Segmentation and Behavioral Analysis},
  author    = {Chen, Xi and Zhai, Hao and Liu, Danqian and Li, Weifu and Ding, Chaoyue and Xie, Qiwei and Han, Hua},
  booktitle = {Proceedings of the Twenty-Ninth International Joint Conference on
               Artificial Intelligence, {IJCAI-20}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},             
  pages     = {5300--5302},
  year      = {2020},
  month     = {7},
  doi       = {10.24963/ijcai.2020/776},
  url       = {https://doi.org/10.24963/ijcai.2020/776},
}

References

@article{A_Common_Hub,
  title={A common hub for sleep and motor control in the substantia nigra},
  author={Liu, Danqian and Li, Weifu and Ma, Chenyan and Zheng, Weitong and Yao, Yuanyuan and Tso, Chak Foon and Zhong, Peng and Chen, Xi and Song, Jun Ho and Choi, Woochul and others},
  journal={Science},
  volume={367},
  number={6476},
  pages={440--445},
  year={2020},
  publisher={American Association for the Advancement of Science}
}

About

SiamBOMB: Siamese network using Background information for real-time Online Multi-species home-cage animal Behavioral analysis. (Updating)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 96.8%
  • Cuda 1.9%
  • C++ 1.1%
  • Other 0.2%