Skip to content

DesignAutomationLaboratory/edocr2

 
 

Repository files navigation

eDOCr2

A tool for performing segmentation and OCR on engineering drawings, primarily focused on mechanical or production drawings.

Installation

Detailed installation steps can be found here.

How to Use

For quick testing, run the test_drawing.py file after downloading the recognizer models from Releases.

Other files are provided for additional functionality:

  • test_train.py: For training custom synthetic recognizers or detectors.
  • test_all.py: For benchmarking against all drawings available in the /tests folder.
  • test_llm.py: For integration with language models (LLMs).

For more detailed information about using these files, refer to the Examples documentation.

Citation

If you find this project useful in your research, please consider citing our paper (preprint):
http://dx.doi.org/10.2139/ssrn.5045921

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%