This is the official GitHub Repo for the paper Robust Multiview Multimodal Driver Monitoring System Using Masked Multi-Head Self-Attention accepted by the MULA workshop at CVPR 2023.
TODO:
- Upload our manually annotated labels.
- Upload the code of MHSA.
- Upload the code of SuMoCo.
- Upload the code for training and evaluation.
- Visit the official DAD repo.
- Find the link for data downloading.
- Download the data.
- Unzip the data into the folder
data
(create if not exists) under this repo.
- Copy and paste
label.csv
into the folderdata
. - Run
python preprocess.py
to generate the pickle files for the training set and the test set. You can specify the arguments to create your own dataset.
- Run
train.py
to train the model. You can specify the arguments, e.g., which data sources or fusion methods to use...