Tesseract 5.4.0 text recognition system now available for release

Published release of optical text recognition system Tesseract 5.4.0, supporting recognition of UTF-8 characters and texts in more than 100 languages, including Russian, Kazakh, Belarusian and Ukrainian. The result can be saved in plain text or in HTML (hOCR), ALTO (XML), PDF and TSV formats. The system was originally created in 1985-1995 in the Hewlett Packard laboratory; in 2005, the code was opened under the Apache license and was subsequently developed with the participation of Google employees. Project sources spread licensed under Apache 2.0.

Tesseract includes a console utility and the libtesseract library for embedding OCR functionality into other applications. From third parties that support Tesseract GUI interfaces you can note gImageReader, VietOCR And YAGF. Two recognition engines are offered: a classic one that recognizes text at the level of individual character patterns, and a new one based on the use of a machine learning system based on an LSTM recurrent neural network, optimized for recognizing entire strings and allowing for a significant increase in accuracy. Ready-made trained models are published for 123 languages. To optimize performance, modules using OpenMP and SIMD instructions AVX2, AVX, AVX512F, NEON or SSE4.1 are offered.

Basic improvements:

  • Added support for rendering and exporting in the format PAGE-XML.
  • It is now possible to train the model using PNG files instead of LSTMF files.
  • Improved rendering to PDF format.
  • The API for determining the tilt of text has been expanded.
  • Fixed performance issues identified when scanning in the Coverity system.

Thanks for reading: