Calligraphic character recognition method based on similarity-sensitive coding

Abstract

The invention discloses a calligraphy character writing style identification method based on similarity sensitivity encoding. The method comprises the following steps of: firstly, collecting calligraphy character images, forming a calligraphy character writing style library according to the writing style types, then, training a similarity sensitivity encoding function by the calligraphy character writing style library, encoding each calligraphy character in the writing style library, and converting the calligraphy character into a binary feature vector; when identifying the writing style, extracting features of the to-be-identified calligraphy character image, and then encoding the features to get the binary feature vector of the to-be-identified calligraphy character image, then, matching the binary feature vector with the encoded writing style library, and taking the writing style represented by the writing style library most similar as the to-be-identified calligraphy character as the identification result. The method solves the defect that the calligraphy character writing style is difficult to define, and provides an effective method capable of identifying the writing style of the calligraphy character better.

Description

technical field [0001] The invention relates to the fields of digital library, image processing and machine learning, in particular to a method for recognizing calligraphy characters based on similarity-sensitive coding. Background technique [0002] Chinese calligraphy works are an important part of Chinese traditional culture. At present, more and more excellent calligraphy works have been digitized and stored in digital libraries. Users often hope to appreciate calligraphy works of the same style at the same time, and many users hope to obtain the style of a certain calligraphy work. However, due to some characteristics of calligraphy characters, the current calligraphy character recognition technology cannot recognize the style of calligraphy characters very well, which makes it difficult to automatically classify calligraphy works according to the style of calligraphy. Contents of the invention [0003] The invention overcomes the shortcoming that it is difficult to...

AI Technical Summary

Problems solved by technology

However, due to some characteristics of calligraphy characters, the current calligraphy character recognition technology cannot recognize the calligraphy characters well, which makes it difficult to automatically classify calligraphy works according to the calligraphy characters.

Images