anl13 / MAMMAL_evaluation

MAMMAL Evaluation

This repository presents how to produce evaluation figures (Fig. 3) proposed in MAMMAL.

How to run the codes

These codes are written in Python and work under both windows 10 and Ubuntu (>=16.04).

1. Install conda. (follow official anaconda tutorial)
2. Create conda virtual environment and install required packages.

conda create -n MAMMAL_eval python=3.7
conda activate MAMMAL_eval
pip install -r requirements.txt

3. Unzip data.zip and the file structure should looks like

MAMMAL_evaluation/
`-- data
    |-- iou_for_eval/
    |-- keypoints_for_eval/
    |-- traj/
`-- draw_Fig2b.py
`-- ...

If data/ folder already exists, then you can simply pass this step. 4. Run drawing codes by

python draw_Fig3b.py
python draw_Fig3c.py
python draw_Fig3e.py
python draw_Fig3f.py
python draw_Fig3g.py
python draw_Fig3h.py
python draw_Fig3i.py
python draw_Fig3j.py

and you will get two output folders: (1) data_for_fig/ saves the middle data; (2) figs/ saves the output figures same to our paper.

If you want to run draw_Fig3i.py, you need to prepare pig_render like

MAMMAL_evaluation/
`-- data
    |-- iou_for_eval/
    |-- keypoints_for_eval/
    |-- traj/
`-- pig_render/
`-- draw_Fig2b.py
`-- ...

pig_render can be downloaded from https://github.com/anl13/pig_renderer.

Citation

If you use these datasets in your research, please cite the paper

@article{MAMMAL, 
    author = {An, Liang and Ren, Jilong and Yu, Tao and Hai, Tang and Jia, Yichang and Liu, Yebin},
    title = {Three-dimensional surface motion capture of multiple freely moving pigs using MAMMAL},
    booktitle = {},
    month = {July},
    year = {2022}
}