This program turns a page photograph into a letter-by-letter transcription. It learned from traced lines and was tested on twenty pages it had never seen, spanning Davis's five proposed scribe labels. It finds and cuts out each line itself. Below, compare the scan, the program's text, and the traced letters. Red marks a wrong letter; · marks one left out. The letters use EVA, a common way to write the Voynich script in a regular alphabet.
Loading…
By Davis's scribe labels
A page, line by line
Other photographs of the same pages
A 2014 imaging project photographed five of these pages again. The program reads those color photographs about as well as Yale's. Images processed to reveal faint ink give it more trouble.