by naver-clova-ocr
Open source · 33k downloads · 19 likes
BROS (BERT Relying On Spatiality) is a pre-trained language model designed to extract key information from documents by combining text with spatial layout. By analyzing OCR (optical character recognition) results that include word coordinates, it excels at tasks like extracting structured lists or specific data from receipts, invoices, or forms. Its unique approach, which incorporates the position of textual elements, significantly improves accuracy compared to traditional text-only models. Ideal for automating the processing of complex documents, it is particularly well-suited to environments where layout is essential to comprehension. The model stands out for its ability to understand the hierarchy and spatial relationships between information, providing a robust solution for structured data extraction.
GitHub: https://github.com/clovaai/bros
BROS (BERT Relying On Spatiality) is a pre-trained language model focusing on text and layout for better key information extraction from documents.
Given the OCR results of the document image, which are text and bounding box pairs, it can perform various key information extraction tasks, such as extracting an ordered item list from receipts.
For more details, please refer to our paper:
BROS: A Pre-trained Language Model Focusing on Text and Layout for Better Key Information Extraction from Documents
Teakgyu Hong, Donghyun Kim, Mingi Ji, Wonseok Hwang, Daehyun Nam, Sungrae Park
AAAI 2022 - Main Technical Track
| name | # params | Hugging Face - Models |
|---|---|---|
| bros-base-uncased (this) | < 110M | naver-clova-ocr/bros-base-uncased |
| bros-large-uncased | < 340M | naver-clova-ocr/bros-large-uncased |