Score: 1

PubMed-OCR: PMC Open Access OCR Annotations

Published: January 16, 2026 | arXiv ID: 2601.11425v1

By: Hunter Heidenreich , Yosheb Getachew , Olivia Dinica and more

Potential Business Impact:

Helps computers read science papers accurately.

Business Areas:
Image Recognition Data and Analytics, Software

PubMed-OCR is an OCR-centric corpus of scientific articles derived from PubMed Central Open Access PDFs. Each page image is annotated with Google Cloud Vision and released in a compact JSON schema with word-, line-, and paragraph-level bounding boxes. The corpus spans 209.5K articles (1.5M pages; ~1.3B words) and supports layout-aware modeling, coordinate-grounded QA, and evaluation of OCR-dependent pipelines. We analyze corpus characteristics (e.g., journal coverage and detected layout features) and discuss limitations, including reliance on a single OCR engine and heuristic line reconstruction. We release the data and schema to facilitate downstream research and invite extensions.

Page Count
14 pages

Category
Computer Science:
CV and Pattern Recognition