Brush Stroke Extraction Based On BP Neural Network

Calligraphy is China's traditional culture and unique forms of artistic expression.It is a very significant task to display and express the beauty of Chinese calligraphy through computers and to inherit the Chinese calligraphy culture.This article is mainly about the field of computer calligraphy,and proposes an algorithm for extraction and segmentation of calligraphy strokes.The paper firstly summarizes the related research in the field of computer calligraphy.The feature information of calligraphy characters is divided into the outline features,skeleton features,and stroke segment features of calligraphy characters.This paper uses the point-to-boundary orientation distance(PBOD)algorithm to extract and represent the calligraphy.Word stroke segment feature information.Based on this,a stroke extraction algorithm based on BP neural network and CFER(Chinese character corner relation)corner detection is proposed.The algorithm flow is as follows:First of all,it is necessary to perform corresponding preprocessing on the calligraphy character pictures,remove the noise and perform binarization.Then the PBOD algorithm calculates the distance between all the pixels in the calligraphy character to the boundary of the calligraphy characters,takes a distance from each point in 3 degrees,and draws a graph based on the angle and distance.The number of curve peaks represents different areas.The cross region,endpoint region,and mid-stroke region of the stroke are selected.Then,the training of BP neural network is performed.The PBOD data of the basic stroke intersection region is used as a training set.The trained BP neural network is used to optimize the distribution curve of calligraphy characters,and the noise of peaks and valleys in the PBOD curve diagram is eliminated and calculated.Accurate peaks and valleys,and through the recognition and classification of neural networks,the end regions,middle regions,and intersections of calligraphy strokes are obtained.Finally,the CFER detection algorithm is used to detect the nearest boundary corner point from the intersection of the strokes,and cut the intersection of the calligraphy characters by connecting the corresponding corner points to extract the strokes of the calligraphy characters.Finally,an independent calligraphic word stroke with Chinese stroke information was obtained through the BP neural network and CFER extraction algorithm.Experiments show that the stroke extraction algorithm based on BP neural network improves the accuracy of the extraction of calligraphy strokes and saves important stroke segment information at the same time.