Study shows manuscript-level data leakage inflates historical script classifier accuracy

A new arXiv paper evaluates protocols for historical script classification, measuring how testing on other pages of training manuscripts inflates accuracy and examining the role of soft labels.

Read the full story at arXiv cs.CV ↗

Timeline · 1 report

1. 2026-11-11 05:00 · arXiv cs.CV
   Evaluation protocols for historical script classification: the cost of manuscript-level leakage and the role of soft labels