THÖR-MAGNI Act provides actions aligned with motion capture data and gazes from egocentric videos.
├── LICENSE
├── Makefile <- Makefile with commands like `make data` or `make train`
├── README.md <- The top-level README for developers using this project.
├── data
│ ├── external <- Data from third party sources.
│ ├── interim <- Intermediate data that has been transformed.
│ ├── processed <- The final, canonical data sets for modeling.
│ └── raw <- The original, immutable data dump.
│
├── docs <- A default Sphinx project; see sphinx-doc.org for details
│
├── models <- Trained and serialized models, model predictions, or model summaries
│
├── notebooks <- Jupyter notebooks. Naming convention is a number (for ordering),
│ the creator's initials, and a short `-` delimited description, e.g.
│ `1.0-jqp-initial-data-exploration`.
│
├── references <- Data dictionaries, manuals, and all other explanatory materials.
│
├── reports <- Generated analysis as HTML, PDF, LaTeX, etc.
│ └── figures <- Generated graphics and figures to be used in reporting
│
├── requirements.txt <- The requirements file for reproducing the analysis environment, e.g.
│ generated with `pip freeze > requirements.txt`
│
├── setup.py <- makes project pip installable (pip install -e .) so src can be imported
├── src <- Source code for use in this project.
│ ├── __init__.py <- Makes src a Python module
│ │
│ ├── data <- Scripts to download or generate data
│ │ └── make_dataset.py
│ │
│ ├── features <- Scripts to turn raw data into features for modeling
│ │ └── build_features.py
│ │
│ ├── models <- Scripts to train models and then use trained models to make
│ │ │ predictions
│ │ ├── predict_model.py
│ │ └── train_model.py
│ │
│ └── visualization <- Scripts to create exploratory and results oriented visualizations
│ └── visualize.py
│
└── tox.ini <- tox file with settings for running tox; see tox.readthedocs.io
Project based on the cookiecutter data science project template. #cookiecutterdatascience
Install miniconda. Then, you can install all packages required by running:
conda env create -f environment.yml && conda activate thor-magni-actions && pip install -e .
- Jump to running thor-magni-actions modeling section if you want to skip these preprocessing steps.
- Prepare thor-magni-tools.
- Change config file to:
in_path: PATH_TO_CSVs/Scenario_{ID}
out_path: PATH_TO/thor-magni-actions/data/external/thor_magni
preprocessing_type: 3D-best_marker
max_nans_interpolate: 100
options:
resampling_rule: 400ms
average_window: 800ms
Change the config in_path and out_path settings accordingly. In this way, we obtain smoother and consistent trajectories.
- From
thor-magni-tools, run for each scenario directory:python -m thor_magni_tools.run_preprocessing - Check your
data/externaldirectory. - To align actions and trajectory data, run for each preprocessed scenario directory:
unzip data/processed/thor_magni/QTM_frames_actions.zip -d data/processed/thor_magni/ && python -m thor_magni_tools.run_actions_merging --actions_path data/processed/thor_magni/QTM_frames_actions.csv --files_dir outputs/data/thor_magni/Scenario_{ID}/ --out_path data/interim/thor_magni/ - To compute features, run:
python -m thor_magni_actions.features.build_features data/interim/thor_magni data/interim/thor_magni - To create a dataset of fixed-length tracklets, run:
python -m thor_magni_actions.data.make_dataset thor_magni data/interim/thor_magni data/processed/thor_magni
To run the k-fold cross validation with the multi-task learning approach:
python -m thor_magni_actions.data_modeling.runners.k_fold_cv 5 thor_magni_actions/data_modeling/cfgs/mtl_tf.yaml