README.md

metadata

language:
  - en
license: cc-by-4.0
size_categories:
  - 1K<n<10K
task_categories:
  - feature-extraction
pretty_name: FATURA 2 invoices
tags:
  - invoices
  - data extraction
  - invoice
  - FATURA2
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
dataset_info:
  features:
    - name: image
      dtype: image
    - name: ner_tags
      sequence: int64
    - name: bboxes
      sequence:
        sequence: int64
    - name: tokens
      sequence: string
    - name: id
      dtype: string
  splits:
    - name: train
      num_bytes: 411874484.6
      num_examples: 8600
    - name: test
      num_bytes: 60569760.6
      num_examples: 1400
  download_size: 342750666
  dataset_size: 472444245.20000005

The dataset consists of 10000 jpg images with white backgrounds, 10000 jpg images with colored backgrounds (the same colors used in the paper) as well as 3x10000 json annotation files. The images are generated from 50 different templates.

https://zenodo.org/records/10371464

---

dataset_info: features: - name: image dtype: image - name: ner_tags sequence: int64 - name: words sequence: string - name: bboxes sequence: sequence: int64 splits: - name: train num_bytes: 477503369.0 num_examples: 10000 download_size: 342662174 dataset_size: 477503369.0 configs: - config_name: default data_files: - split: train path: data/train-*

@misc{limam2023fatura, title={FATURA: A Multi-Layout Invoice Image Dataset for Document Analysis and Understanding}, author={Mahmoud Limam and Marwa Dhiaf and Yousri Kessentini}, year={2023}, eprint={2311.11856}, archivePrefix={arXiv}, primaryClass={cs.CV} }