MOCHA: Real-Time Motion Characterization via Context Matching

Official implementation of 'MOCHA: Real-Time Motion Characterization via Context Matching'.

Project Page: https://dk-jang.github.io/MOCHA_SIGASIA2023/

The code provides:

• Data preparation
• Pretrained model and Demo
• Trainging code for MOCHA

Installation and Setup

First, clone the repo. Then, we recommend creating a clean conda environment, installing all dependencies, and finally activating the environment, as follows:

git clone https://github.com/DK-Jang/MOCHA_private.git
cd MOCHA_private
conda env create -f environment.yml
conda activate MOCHA

Datasets and pre-trained networks

• To run the demo, please download the bvh, and pre-trained parameters both.

[Recommend] To download the datasets and the pretrained-networks, run the following commands:

bash download.sh datasets
bash download.sh pretrained-network

If you want to generate train datasets from the scratch, run the following commands:

python ./preprocess/generate_database_bin.py    # generate mirrored bvh files

How to run the demo

After downloading the pre-trained parameterss, you can run the demo.
We use the post-processed results (Ours_*.bvh) for all the demo.

To generate motion characterization results, run following commands:

python test_fullframework.py

Generated motions(bvh format) will be placed under ./results.
Src_*.bvh: ground-truth source motion,
Ours_*.bvh: characterized output motion,

Acknowledgements

This repository contains pieces of code from the following repositories:
Learned Motion Matching.
Motion Puzzle: Arbitrary Motion Style Transfer by Body Part.