---
dataset_info:
  features:
  - name: image
    dtype: string
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 3816064970
    num_examples: 7961
  download_size: 3514907019
  dataset_size: 3816064970
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: apache-2.0
language:
- es
tags:
- finance
pretty_name: EnglishOCR
size_categories:
- 10K<n<100K
task_categories:
- image-to-text
---

----------------------------------------------------------------
# Dataset Card for EnglishOCR Dataset


### Dataset Summary
The EnglishOCR dataset contains images derived from regulatory documents from SEC EDGAR company filings. This dataset is used for benchmarkingg and evaluating Large Language Models ability on converting unstructured dcuments, such as pdfs and images, into machine readable format, particularly in finance domain, where the conversion task is more complex and valuable.

### Supported Tasks

- **Task:** Image-to-Text
- **Evaluation Metrics:** ROUGE-1

### Languages

- English
  
## Dataset Structure

### Data Instances
Each instance in the English OCR dataset comprises 2 fields:
- **image** : image of company filings in 2025 (10-K,10-Q), each image represent one page in pdf
- **text**: ground truth of text

### Data Fields
- **image** : string - Base64-encoded png
- **text**: extracted text from pdf files

## Dataset Creation

### Curation Rationale
The EnglishOCR dataset was curated to support research and development on information extraction techniques and layout retain ability for unstructured documents in financial domain in English. By providing real-world company filings in unstructured format with ground truth, the dataset seeks to address challenges in extracting informat as well as layouts of complex documents, and convert into machine-readable format.

### Source Data
#### Initial Data Collection and Normalization
- The source data are company filings from SEC EDGAR system. 
- Subset of company filings in 2025 are downloaded in html format, and converted into correspondig pdf versions.
- The pdf files are splited and converted into images, and text of the images are extracted.
- The extracted text are used for matching html chunk and correct image.

#### Who are the Source Language Producers?
- The source data are regulatory documents from SEC EDGAR system: https://www.sec.gov/search-filings

### Annotations

#### Annotation Process
- The dataset was prepared by collecting, matching, spliting, and converting regulatory documents in English
- No further manual annotations were applied beyond the original curation process.

#### Who are the Annotators?

- The dataset stems from publicly available company filings.
- No external annotation team was involved beyond this.

### Personal and Sensitive Information

- The EnglishOCR dataset does not contain any personally identifiable information (PII) and is strictly focused on English-language regulatory data. No personal or sensitive information is present in the dataset.

## Considerations for Using the Data

### Social Impact of Dataset
This dataset enables AI models to extract structured information from scanned financial documents, promoting transparency and accessibility. By aligning page-level PDF images with accurate ground truth text, it supports the development of fairer, more inclusive models that work across diverse formats and languages. 

### Discussion of Biases
- The source data is limited to company filings, it may underrepresent other financial document types such as tax records, bank statements, or private company reports, potentially limiting model generalizability.

### Other Known Limitations
- The matching process is based on similarity scores between OCR extracted text from image and html chunk, which may cause inaccuracies, or lost of information in matching process, potentially affecting training quality and evaluation reliability.
- While the dataset covers company filings, it may lack sufficient variety in layout styles (e.g., handwritten notes, non-standard financial forms, embedded charts), which could limit a model’s ability to generalize to less structured or unconventional financial documents.

## Additional Information

### Dataset Curators
- Yueru He
- The FinAI Team

### Licensing Information

- **License:** Apache License 2.0

### Citation Information

If you use this dataset, please cite:

```bibtex
@misc{peng2025multifinbenmultilingualmultimodaldifficultyaware,
      title={MultiFinBen: A Multilingual, Multimodal, and Difficulty-Aware Benchmark for Financial LLM Evaluation}, 
      author={Xueqing Peng and Lingfei Qian and Yan Wang and Ruoyu Xiang and Yueru He and Yang Ren and Mingyang Jiang and Jeff Zhao and Huan He and Yi Han and Yun Feng and Yuechen Jiang and Yupeng Cao and Haohang Li and Yangyang Yu and Xiaoyu Wang and Penglei Gao and Shengyuan Lin and Keyi Wang and Shanshan Yang and Yilun Zhao and Zhiwei Liu and Peng Lu and Jerry Huang and Suyuchen Wang and Triantafillos Papadopoulos and Polydoros Giannouris and Efstathia Soufleri and Nuo Chen and Guojun Xiong and Zhiyang Deng and Yijia Zhao and Mingquan Lin and Meikang Qiu and Kaleb E Smith and Arman Cohan and Xiao-Yang Liu and Jimin Huang and Alejandro Lopez-Lira and Xi Chen and Junichi Tsujii and Jian-Yun Nie and Sophia Ananiadou and Qianqian Xie},
      year={2025},
      eprint={2506.14028},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2506.14028}, 
}
```