Ten clicks per book beat more data, ResNet-50, and 1024px for Urdu digitization
Ibteda Digital Library, a Pakistani community archive, mined 575,729 crop labels from a decade of Photoshop work by registering finished pages back to raw scans. Scaling training books from 378 to 572, swapping in ResNet-50, and raising input resolution all failed to move held-out pass@80. Per-volume error analysis showed why: the model couldn't learn an invisible operator preference. Ten corrected crops per book, combined via median residual, lifted pass@80 from 0.71 to 0.83.