OCR copes poorly with non-animal things (e.g. plants) #87
Description
In https://pfocr.wikipathways.org/figures/PMC8505746__fpls-12-748204-g001.html, a figure about some protein complexes in plants is incorrectly matched to animal (human, fly) genes. Starting with this problematic analysis, I searched for "photosynthesis" in the database and found many similar problems:
- https://pfocr.wikipathways.org/figures/PMC12109773__cimb-47-00313-g007.html
- https://pfocr.wikipathways.org/figures/PMC12073235__plants-14-01372-g001.html
- https://pfocr.wikipathways.org/figures/PMC11676349__ijms-25-13718-g009a.html
I think some kind of lightweight NLP on the article title, abstract, and/or the figure caption should be done to first determine the taxonomic scope of the figure.