EMAG: Ego-motion Aware and Generalizable 2D Hand Forecasting from Egocentric Videos

[Paper][Supplementary][Project Page][Poster]

This is the official code release for our ECCVW 2024 paper
"EMAG: Ego-motion Aware and Generalizable 2D Hand Forecasting from Egocentric Videos".

System

• OS: Linux / Ubuntu 20.04
• Python Version: Python 3.8.10
• CUDA Version: CUDA 11.7

🔨 Installation

# Create a virtual environment
python3 -m venv emag
source emag/bin/activate

# Install the dependencies
pip install -r requirements.txt

🔥 Training

python3 lit_main.py train=True test=False

🔍 Evaluation

To evaluate the model, please run the following command.

python3 lit_main.py train=False test=True

✍️ Citation

If you use this code for your research, please cite our paper.

@inproceedings{Hatano2024EMAG,
    author = {Hatano, Masashi and Hachiuma, Ryo and Saito, Hideo},
    title = {EMAG: Ego-motion Aware and Generalizable 2D Hand Forecasting from Egocentric Videos},
    booktitle = {European Conference on Computer Vision Workshops (ECCVW)},
    year = {2024},
}