paride92 / layout-parser

A Unified Toolkit for Deep Learning Based Document Image Analysis

A unified toolkit for Deep Learning Based Document Image Analysis

---

What is LayoutParser

LayoutParser aims to provide a wide range of tools that aims to streamline Document Image Analysis (DIA) tasks. Please check the LayoutParser demo video (1 min) or full talk (15 min) for details. And here are some key features:

• LayoutParser provides a rich repository of deep learning models for layout detection as well as a set of unified APIs for using them. For example,

  Perform DL layout detection in 4 lines of code

  import layoutparser as lp
  model = lp.AutoLayoutModel('lp://EfficientDete/PubLayNet')
  # image = Image.open("path/to/image")
  layout = model.detect(image) 

• LayoutParser comes with a set of layout data structures with carefully designed APIs that are optimized for document image analysis tasks. For example,

  Selecting layout/textual elements in the left column of a page

  image_width = image.size[0]
  left_column = lp.Interval(0, image_width/2, axis='x')
  layout.filter_by(left_column, center=True) # select objects in the left column 

  Performing OCR for each detected Layout Region

  ocr_agent = lp.TesseractAgent()
  for layout_region in layout: 
      image_segment = layout_region.crop(image)
      text = ocr_agent.detect(image_segment)

  Flexible APIs for visualizing the detected layouts

  lp.draw_box(image, layout, box_width=1, show_element_id=True, box_alpha=0.25)

  Loading layout data stored in json, csv, and even PDFs

  layout = lp.load_json("path/to/json")
  layout = lp.load_csv("path/to/csv")
  pdf_layout = lp.load_pdf("path/to/pdf")

• LayoutParser is also a open platform that enables the sharing of layout detection models and DIA pipelines among the community.

  Check the LayoutParser open platform
  Submit your models/pipelines to LayoutParser

Installation

After several major updates, layoutparser provides various functionalities and deep learning models from different backends. But it still easy to install layoutparser, and we designed the installation method in a way such that you can choose to install only the needed dependencies for your project:

pip install layoutparser # Install the base layoutparser library with  
pip install "layoutparser[layoutmodels]" # Install DL layout model toolkit 
pip install "layoutparser[ocr]" # Install OCR toolkit

Extra steps are needed if you want to use Detectron2-based models. Please check installation.md for additional details on layoutparser installation.

Examples

We provide a series of examples for to help you start using the layout parser library:

1. Table OCR and Results Parsing: layoutparser can be used for conveniently OCR documents and convert the output in to structured data.

2. Deep Layout Parsing Example: With the help of Deep Learning, layoutparser supports the analysis very complex documents and processing of the hierarchical structure in the layouts.

Contributing

We encourage you to contribute to Layout Parser! Please check out the Contributing guidelines for guidelines about how to proceed. Join us!

Citing layoutparser

If you find layoutparser helpful to your work, please consider citing our tool and paper using the following BibTeX entry.

@article{shen2021layoutparser,
  title={LayoutParser: A Unified Toolkit for Deep Learning Based Document Image Analysis},
  author={Shen, Zejiang and Zhang, Ruochen and Dell, Melissa and Lee, Benjamin Charles Germain and Carlson, Jacob and Li, Weining},
  journal={arXiv preprint arXiv:2103.15348},
  year={2021}
}