Default highlights (blue rectangles) misplaced over names?
Many highlights are already over the images, but very often they seem to be slightly misplaced, even covering words on other lines.
I wonder if this is causing the OCR to be less accurate.
In this example, D[oñ]a is the word to be identified. Interface offers two different highlights by default:
The first one, highlight covers D[oña] but also Manuel above.
In the second one, highlight only covers a bit of the D, and overextends below to Ra-.
Could this be related to problems with the OCR?
Would it be helpful if I create a new highlight? something like
Image comes from
"San Juan Bautista, Guacarí, Valle del Cauca, Colombia records," images, FamilySearch (https://www.familysearch.org/ark:/61903/3:1:3QS7-L9VH-PG8M?view=explore : Mar 1, 2026), image 403 of 595; Archivo Diocesano de Buga (Colombia).
Image Group Number: 004265911
Thank you!
Answers
-
@JuanZuluaga3 In full name review, you can fix this problem - Step 1 Is the full name highlighted NO. Then in Step 2 Click on the yellow existing highlight parameters (boundaries) and minimize the size of the highlight so it only includes the correct word/name then go on to complete step 3 — the name will come up without the extra unwanted words and you can complete step 4 and hit done
0 -
Than you Mary Ellen,
these are hundreds of records, but not in Full Name Review. I suppose that by me selecting the highlight, I'm helping to train the AI. If not, then I should not bother changing the highlight, and go on to other task.
0
