FIELD: physics.
SUBSTANCE: invention relates to systems and methods of character recognition using artificial intelligence. Such a result is achieved due to that the method includes obtaining a text image, wherein the text on the image contains one or more words in one or more sentences; obtaining a text image as first initial data for a set of trained models of machine learning, storing information on compatibility of words and frequency of their combined use in real sentences; obtaining one or more final output data from a set of trained machine learning models, as well as extracting from one or more final output data of one or more assumed sentences from text on an image. Each of one or more assumed sentences contains possible sequences of words.
EFFECT: high efficiency of recognizing text by using a set of models of machine learning, which enable analysis of context of words of text on an image with high quality.
21 cl, 27 dwg
