Tesseract: List of parameters considered by Tesseract 4

Created on 22 Jan 2018  路  4Comments  路  Source: tesseract-ocr/tesseract

Hello,
is the a (comprehensive) list of settings which work in Tesseract 4?
I can only find a list for Tesseract 3 and most of them have no effect in v4.

For example I want to modify the minimum confidence value for a word to be detected.

Kind regards

Most helpful comment

As a follow up, @amitdo from the top of your head what parameters do you think are useful with lstm? I'm trying to improve some results on tables and I'm having a hard time finding out what parameters work (so far, all the ones that I recall were useful for v3 don't do anything)

All 4 comments

tesseract --print-parameters

But many of them does not have any effect on the new lstm based ocr engine.

thanks for the command.
I assume a list of v4 paramters does not exists yet.

As a follow up, @amitdo from the top of your head what parameters do you think are useful with lstm? I'm trying to improve some results on tables and I'm having a hard time finding out what parameters work (so far, all the ones that I recall were useful for v3 don't do anything)

Most parameters related to the layout analysis should work as in 3.05.

For tables maybe the info in #825 will help you.

For the text recognition (ocr), I can clearly see that most parameters are related to the legacy engine.
I don't have a list of the parameters that affect the lstm based ocr engine.

Was this page helpful?
0 / 5 - 0 ratings

Related issues

clarkk picture clarkk  路  7Comments

clarkk picture clarkk  路  6Comments

Shreeshrii picture Shreeshrii  路  4Comments

clarkk picture clarkk  路  3Comments

ivder picture ivder  路  7Comments