A computer vision-based approach for automatically extracting data from bar chart raster images to facilitate meta-analysis of biomedical literature
Alexander Cardaras, Seungjun Kim, Ye Yuan, Itamar Livnat, Ryan T. Yanagihara, Rosita Saul, Gabrielle Montes De Oca, Kai Zheng, Andrew W. Browne · 2026
Read A computer vision-based approach for automatically extracting data from bar chart raster images to facilitate meta-analysis of biomedical literature by Alexander Cardaras, Seungjun Kim, Ye Yuan, Itamar Livnat, Ryan T. Yanagihara, Rosita Saul, Gabrielle Montes De Oca, Kai Zheng, Andrew W. Browne (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13426920.txt.
Description
Cardaras and colleagues published a 2026 study on software that reads bar charts in papers. Bar charts hold key trial results, yet their data sit locked in pictures. Review teams copying numbers by hand waste hours and risk slips. The team built a four-step pipeline that pulls numbers out automatically.
The pipeline sorts chart parts, reads labels, and rebuilds values from the axes. On 28 charts from eye-disease trials, it spotted most chart types correctly. About four in five rebuilt values landed within five percent of hand-checked truth. Most errors traced to misread digits or fuzzy bar edges in scans. The proof-of-concept tool could speed up large evidence reviews greatly. Hand checks used a semi-automated labeling tool as reference standard.
Catalog record
| CREATOR | Alexander Cardaras, Seungjun Kim, Ye Yuan, Itamar Livnat, Ryan T. Yanagihara, Rosita Saul, Gabrielle Montes De Oca, Kai Zheng, Andrew W. Browne |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35 KB |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |