Layout Analysis of Scanned Documents

Document Layout Analysis using YOLOv8
View Demo · Report Bug · Request Feature

Table of Contents

1. Updates
2. About The Project
   • Built With
3. Getting Started
   • Prerequisites
   • Installation
4. Works Cited
5. Acknowledgments

Updates

In this project, I provided 1 object detection model trained on the existing YOLOv8 weights. They are uploaded in my Hugging Face Space of the project. If you feel the need to use or fine-tune the models in any parts of your work, please cite this repository. Thank you, and don't forget to give this repo a 🌟!

About The Project

Due to the lack of computational resources, I only performed the training process on the Doclaynet-base dataset which contains 6910 train images, 648 val images, 499 test images. However, the model could perform relatively well, further proving the superiority of YOLOv8 model.

(back to top)

Built With

• YOLOv8
• Hugging Face Spaces
• Gradio
• Google Colab

(back to top)

Prerequisites

1. python 3
2. ultralytics
3. numpy
4. opencv-python

Installation

1. Clone the repo

   git clone https://github.com/LynnHaDo/Document-Layout-Analysis.git

2. Install packages

   pip install ultralytics
   pip install numpy
   pip install opencv-python

3. Download Doclaynet dataset and save it as datasets/doclaynet-base
4. (Optional) Download pretrained YOLOv8s weights

(back to top)

Works Cited

1. Ultralytics YOLOv8

   authors:
    - family-names: Jocher
      given-names: Glenn
      orcid: "https://orcid.org/0000-0001-5950-6979"
    - family-names: Chaurasia
      given-names: Ayush
      orcid: "https://orcid.org/0000-0002-7603-6750"
    - family-names: Qiu
      given-names: Jing
      orcid: "https://orcid.org/0000-0003-3783-7069"
   title: "YOLO by Ultralytics"
   version: 8.0.0
   date-released: 2023-1-10
   license: AGPL-3.0
   url: "https://github.com/ultralytics/ultralytics"

2. Doclaynet-base dataset

   @article{doclaynet2022,
    title = {DocLayNet: A Large Human-Annotated Dataset for Document-Layout Segmentation},
    doi = {10.1145/3534678.353904},
    url = {https://doi.org/10.1145/3534678.3539043},
    author = {Pfitzmann, Birgit and Auer, Christoph and Dolfi, Michele and Nassar, Ahmed S and Staar, Peter W J},
    year = {2022},
    isbn = {9781450393850},
    publisher = {Association for Computing Machinery},
    address = {New York, NY, USA},
    booktitle = {Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining},
    pages = {3743–3751},
    numpages = {9},
    location = {Washington DC, USA},
    series = {KDD '22}
    }

Contact

Linh Do - [email protected]/[email protected] (personal)

Project Link: https://github.com/LynnHaDo/Document-Layout-Analysis

LinkedIn: https://linkedin.com/in/Linh Do

(back to top)