Efficient pipelines and models for document AI, with Docling support via plugins.
With uv, add to your project using:
uv add cells2table| Optional | Description |
|---|---|
huggingface |
For downloading models from HF |
onnxruntime |
For using ONNX Runtime as runtime |
transformers |
For using transformers as runtime |
opencv-headless |
For using ONNX models |
opencv |
For using ONNX models and the CLI |
Check the documentation.
@InProceedings{Pastorello2027TableStructureRecognition,
author = {Pastorello, Jo{\~a}o Speranza and
Tamiosso, Gustavo Lopes and
de Oliveira, Gabriel Bauer and
Geiger, Albert and
da Silveira, Evandro Dalbosco and
Balreira, Dennis Giovani and
Tavares, Anderson Rocha},
editor = {Recamonde-Mendoza, Mariana and Barros, Rodrigo C.},
title = {Efficient Table Structure Recognition via Deep Learning and Lightweight Heuristics},
booktitle = {Intelligent Systems},
year = {2027},
publisher = {Springer Nature Switzerland},
address = {Cham},
pages = {643--657},
isbn = {978-3-032-39895-6},
doi = {10.1007/978-3-032-39895-6_43}
}