Add simple orientation detection #34
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
After some local tests I think an alternative approach might be to:
|
Tesseract's has built-in script and orientation detection but it is part of the classic (pre-LSTM) engine, which has been compiled out to reduce binary size. Hence this initial implementation use Leptonica's more simplistic orientation detection, which is based on counting numbers of ascenders and descenders, as described on pages 12-14 of http://www.leptonica.org/papers/skew-measurement.pdf.
In adding this I encountered issues where the same rotated image dropped into Safari, Chrome and Firefox could give different results. I believe this has to do with how the EXIF rotation information is handled by the various browser APIs used to load and draw images, but this still needs to be debugged.
The confidence value is currently 0 if an error occurred or 1 otherwise. This at least creates a space in the API to include a confidence score in the result.
robertknight
force-pushed
the
orient-detect
branch
from
3803c2f
to
26893cc
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Add simple orientation detection using Leptonica's
pixOrientDetectfunction. This was used in Tesseract because Tesseract's implementation requires the legacy (non-LSTM) engine, which is not compiled in. Leptonica's algorithm relies mostly on "the preponderence of ascenders over descenders in languages with roman characters", per this paper. Tesseract's approach which is not being used is described here.TODO:
getOrientationAPI to indicate uncertainty in the result or errors in the process. Currently it returns 0 in the event of any error, and has no way to represent confidence in the result.