Method of automatic language identification for multi-lingual text recognition

Abstract

The disclosed invention utilizes a complex estimation-based approach to identify languages of portions of a multi-lingual text, recognized from a bit-mapped image. The method comprises besides the traditional steps like the document segmentation, new ones such as generating and testing of a hypothesis about the characters in the word tokens. The method further includes definition of selected language models set, word estimation via language models, dictionaries set definition for language selection, estimation of word correspondence with chosen languages, calculating a complex estimation for the word taking into account the most or all of above mentioned estimations. The complex estimation may also include factor of characters and/or words mutual correspondence within the line and/or the text, mutual geometric correspondence of characters within the word and/or the line, linguistic correspondence of the word with neighbors, estimation of image of word token reconstruction accuracy in the presence of distortion.

Dependencies are inferred from the claim text.