Research On Automatic Scoring Method Of Chalk Calligraphy Based On Template Matching

Under the background of the wave of information technology,students’ writing ability has declined significantly,and the phenomenon of "forgetting to write" is not uncommon.The writing of chalk Chinese characters,as an essential quality and basic cultivation of normal school students,should be highly valued.In practical teaching,there is a shortage of professional teachers and guidance in the practice of chalk Chinese characters.Based on the idea of template matching,the thesis explores the standard scoring method for chalk-oriented Chinese character writing,aiming at providing useful thoughts and suggestions for solving the above problems.The research work is mainly divided into the following aspects:1.Construction of chalk Chinese character template dictionary.At present,the scoring standard of chalk Chinese characters has not been unified,and there is no theoretical basis for using computer aided chalk Chinese characters to score.In view of the above problems,by consulting a large number of literature and combining with the guidance of relevant experts,the writing standards of chalk characters are studied,the characteristics affecting the score of chalk characters are summarized and processed quantitatively,a standard chalk character scoring system is formed,and the construction of a chalk character template dictionary is completed.2.Feature extraction of chalk Chinese characters.Feature extraction of chalk Chinese characters is an important step in chalk Chinese character scoring.It can obtain effective digital information from chalk Chinese characters images that can represent the writing characteristics of chalk Chinese characters.The feature extraction of powder characters is divided into three levels: whole characters,components and strokes.At the whole characters level,geometric knowledge and function formula are used to extract traditional features such as center of gravity and skeleton.At the same time,VGG16 model is used to explore the deep characteristics of different Chinese characters,as an effective supplement to the traditional method.In view of the lack of research on component extraction of off-line handwritten Chinese characters and the difficulty in solving the problem of component adhesion,a component extraction method based on connected region merging is proposed.Meanwhile,the traditional dripping algorithm is improved to cut the cohesive components,so as to realize the extraction of the size and aspect ratio of the chalk character components.At the stroke level,the method based on local information is used to complete the extraction of stroke strokes,and the characteristic information such as stroke endpoint and length is calculated,which lays the foundation for the subsequent calligraphy evaluation of chalk Chinese characters.3.Study on automatic scoring method of chalk Chinese characters.Firstly,the Kuhn-Munkres algorithm is used to match the features of the students’ penned characters with those in the template dictionary.Secondly,according to the characteristics of different data types,the corresponding method is adopted to carry out similarity calculation.Thirdly,multiple linear regression method and BP neural network are used to construct the automatic scoring model of chalk Chinese calligraphy Finally,by comparing the results of the two methods above,a better method is chosen to achieve the task of the chalk Chinese characters scoring.The experimental results show that the method based on BP neural network model is better than the method based on multiple linear regression,and has good accuracy.It can make beneficial exploration for the computer aided chalk Chinese character scoring research.