- Python ^3.8 and < 3.11
- (Optional) Poetry for dependencies
- (Optional) Git LFS so training data. (See Git Large File Storage
For development
- (Optionally) Docker
- (Optionally) Vscode with devcontainer extension
pip install -r requirements.txt
# or using poetry
poetry install
poetry shell
# optionally to get the files (not needed for inference)
git lfs pull
NOTE: Print help using python main.py --help
IMPORTANT. This program only works well on binarized images.
The main command will read all .jpg images from INPUT-DIRCTORY and output the .txt files in the results folder. To do this, execute the following command from the root of the project.
python main.py run [INPUT-DIRECTORY]
For example:
python main.py run test-images
The output directory can be changed using the --o flag. Defaults to results.
python main.py run test-images --o out/folder
To only select specific images it is possible to use the --suffix flag. This will only select .jpg files that end with a specific string. Usefull for selecting only binarized images.
python main.py run test-images --suffix binarized
Not usefull for evaluation, but other commands can be used to train, segment and perform augmentation on the dataset.
See the full list of commands using:
python main.py --help
Get instructions for a specific command using
python main.py [command] --help
# for example
python main.py run --help
- Prepare: prepares the data by unzipping the data in the correct folder
- Augment: Augments the data with different augmentation methods, (dilate, )
- Train: Trains the classifier on the AUGMENTED dataset
- Run: the model on the image-data
- Linesegment Segments into lines
- Charactersegment - Segments the lines into characters
Development is easiest inside a devcontainer, and we strongly recommend to use it. Install the extension and open the project inside a container. Then execute the following commands.
poetry config virtualenvs.in-project true
poetry install
Activate the environment and execute the program by.
poetry shell
python main.py