From 19a23907396db560375da0fa913ad8870a3d65f2 Mon Sep 17 00:00:00 2001 From: David Ragnar Nelson <35697532+drnelson6@users.noreply.github.com> Date: Fri, 31 May 2024 11:36:09 -0400 Subject: [PATCH] Add instructions for correcting line masks --- transcription.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/transcription.md b/transcription.md index 27787ca..5b8ae0b 100644 --- a/transcription.md +++ b/transcription.md @@ -63,6 +63,8 @@ Next, delete any pages that do not contain text. If you are unsure whether a pag - Correct any lines that are too small or too large. - Delete and extraneous lines. - Manually add any lines that the automatic process missed. + - Verify line masks: does a line mask exist? Does it reasonably cover the text? + - Note: do **NOT** try to manually adjust masks. Instead, fiddle with the baseline until the mask reasonably encompasses the text. If the mask is truly egregious, ask a project manager for assistance. - Manually segment any pages that cannot be automatically segmented. Please see the [official documentation](https://escriptorium.readthedocs.io/en/latest/segment/#text-line-segmentation) to learn how the segmentation tool operates.