This package represents the Automatic Video Analysis Framework (AVAF) for the H2020 project Visual History of the Holocaust. It includes various video analysis techniques (e.g. Shot Boundary Detection, Shot Type Classfication and Camera Movements Classification).
PDF format: vhh_core_pdf
HTML format (only usable if repository is available in local storage): vhh_core_html
Requirements:
- Ubuntu 18.04 LTS
- python version 3.6.x
Create a virtual environment:
- create a folder to a specified path (e.g. /xxx/vhh_core/)
python3 -m venv /xxx/vhh_core/
Activate the environment:
source /xxx/vhh_core/bin/activate
Checkout vhh_core repository to a specified folder:
git clone https://github.com/dahe-cvl/vhh_core
Install dependencies and plugins
cd /xxx/vhh_core/pip install -r requirements.txt
Install PyTorch
Install a Version of PyTorch depending on your setup. Consult the PyTorch website for detailed instructions.
Setup environment variables:
- Activate environment:
source /xxx/vhh_core/bin/activate - Let PATH_TO_DIR be the path to the directory containing this repository. Change pythonpath:
export PYTHONPATH=$PYTHONPATH:/PATH_TO_DIR/vhh_core/:/PATH_TO_DIR/vhh_core/Develop/:/PATH_TO_DIR/vhh_core/Demo/
Import Models
Executing the program for the first time (see next section) will create the desired folder structure. Insert the models for SBD and STC into the corresponding subdirectories of /models.
Folders for the downloaded videos and the results will be automatically created in the second run.
Run demo script
- change to root directory of the repository
python Demo/run_automatic_annotation_process.py- Import models for SBD and STC.
python Demo/run_automatic_annotation_process.py