v0.7-0

Bug fixes:

• data file downloading migrated to Tesseract OCR Github repository, fixes the main issue which blocked the usage of the app
• segmentation fault on application exit fixed (occured only if the app was closed without running OCR)
• some minor UI fixes for better looks on new devices

Features / other stuff:

• x86 version released to the Jolla store (for Jolla tablet)
• Tesseract updated to 3.05, Leptonica to 1.74
  • language files have to be downloaded again due to version and language data remote location update
• new languages, total count now 101 + (equation detection)

v0.6-0

Bug fixes:

• limit max tile size to 1000
• fix for too early firing signal in PDF file preview page

EDIT: Added i486 binary. Version is identical to armv7hl version. Tested only on emulator. Only portrait orientation supported. May look a bit crappy on the tablet until UI is adapted for the tablet.

Textractor 0.5 pre-release

This is a pre-release for public testing before I roll out the actual release to the Jolla Store. This release includes multiple new features, new dependencies on third party libs and a lot of new code so it is possible there are some bugs.

Changelog:

New features:

• Camera orientation modes: automatic, landscape, portrait
• Cropping with perspective correction
• Reordering in language selection dialog: already downloaded languages are on top of the list
• Text extraction from PDF files
• Preprocessed image now shows up at the processing page (known issue: some images show up completely black but in reality they are ok)
• OCR status in application cover

Bug fixes:

• Increased scaling factor for small images to prevent preprocessing error during background normalization
• SFOS 2.0 specific: replace missing icons with available ones

Miscellaneous:

• Update Tesseract OCR from 3.04rc1 to 3.04.00 release
• Poll tesseract monitor faster (500 ms -> 250 ms)
• Fixed some typos and updated hints page
• Small UI fixes and changes

v0.4-0

Bugfixes:

• #6 (partial fix)
• #7
• #8

First proper release

This is first "proper" release.

See documentation for general usage instructions: http://skvark.github.io/Textractor/

Pre-release 0.2

Bugfixes:

• preprocessing failed on small images
• image rotation works now correctly for orientation in which the usb port points to right
• do not overwrite images selected via gallery - create copy instead and modify it

Textractor alpha release

This is the first public release of Textractor. Please keep in mind that Textractor is still in alpha state so there're probably bugs and features are limited.

Bug reports and feature requests can be submitted to the issues of this repository.

Known issues/missing features etc.:

• Only images taken with Jolla's camera are confirmed to work!
• Recognition can't be canceled due to possible memory leaks in Tesseract OCR (I have to investigate this further)
• Preprocessed image is always overwritten in new recognition event
• Image can't be cropped before recognition
• Settings are pretty complex and some parameters are missing from there (detailed explanations will be written at some point)
• Icon is the default Sailfish one, hopefully someone will make better one :)

General info:

• Language file sizes vary between 0,5 - 10 MB (they are not shipped with the application because total size would be then over 200 MB)
• results are filtered with a hardcoded confidence limit of 20 (will be added to settings)