Dawsonia#
Digitize hAndWritten obServatiONs In weather journAls
Dawsonia is a young project aimed at data-rescue of weather journals. It specializes in digitization of hand-written numeric data in the form of tables. We aim to use a combination of
image processing
machine learning
to achieve this. The digitization pipeline is implemented in Python, using well-known open-source scientific libraries.
- User guide
- API Reference
- Wishlist
- Challenges for Hackathon 2024
- Improve table detection method
OPENCV_CONTOURS
- Active learning through semi-supervised training of the HTR model
- Use image-processing based table detection to train a AI-based table detection
- Extend training dataset using public datasets and make it compatible with Dawsonia
- Extend training dataset using different image augmentation methods
- Choose an image augmentation library to and write a script to generate
- Create a dashboard for digitization and deploy it to Hugging Face
- Improve table detection method