Chinese character image stroke extraction method and system based on full convolutional neural network

Abstract

The invention belongs to the field of computer vision and pattern recognition, particularly relates to a Chinese character image stroke extraction method and system based on a full convolutional neural network, and aims to solve the problem that handwriting character strokes written freely are difficult to extract. The Chinese character image stroke extraction method comprises the following steps:carrying out region extraction on an acquired Chinese character image; carrying out skeletonization operation on the overlapped area and the non-overlapped area; calculating the coherence between anystroke segments of the overlapped area after skeletonization; and connecting the stroke segments belonging to the same stroke in the overlapped area, and combining the stroke segments directly connected in the non-overlapped area into a complete skeleton form stroke. On one hand, stroke extraction of handwritten Chinese characters can still be achieved under the condition that strokes of the handwritten Chinese characters which are written freely are overlapped; and on the other hand, a character synthesis method is adopted for obtaining a training sample, and different marking information ofthe training sample in different tasks is attached, so that the labor cost is greatly saved.

Description

technical field [0001] The invention belongs to the field of computer vision and pattern recognition, and in particular relates to a method and system for extracting strokes of Chinese character images based on a fully convolutional neural network. Background technique [0002] Stroke extraction from Chinese character images plays an important role in text recognition research and related applications based on structural analysis. Single-character classification of Chinese handwritten / printed characters based on deep learning technology has achieved a fairly high accuracy rate. However, in many applications, people not only care about the classification of characters, but also pay attention to stroke interpretation, writing quality evaluation, shape beautification, and font design. And other problems, and this requires segmentation and extraction of strokes in text images. [0003] For the problem of stroke extraction of offline Chinese characters, there are mainly two type...

AI Technical Summary

Problems solved by technology

[0013] In order to solve the above-mentioned problems in the prior art, that is, the difficult problem of free-writing handwritten character stroke extraction, the present invention provides a Chinese character image stroke extraction method based on a fully convolutional neural network. The extraction method includes:

Images