Project page: scam.silviacascianelli.com
SCAM is a dataset of handwritten Sahidic Coptic manuscript pages, with manually annotated text-line polygons and HTR transcriptions. It is split into two subsets:
Subset
Source
Pages
Lines
SCAM-A
Biblioteca Apostolica Vaticana
24
1,445
SCAM-B
Biblioteca Nazionale di Napoli
15
1,795
Full
Combined
39
3,240