Font Size: a A A

Machine Learning-based Calligraphy Character Recognition And Intelligent Judgment

Posted on:2021-05-16Degree:MasterType:Thesis
Country:ChinaCandidate:R T ShaoFull Text:PDF
GTID:2415330629986081Subject:Control engineering
Abstract/Summary:PDF Full Text Request
In today's Internet information age,research in calligraphy education is still in its infancy.The main purpose of this article is to help beginners in calligraphy to better understand the effects of their copy works.After consulting a large number of relevant materials and fully communicating with Teacher Zhang Kejiang,I found some common problems for calligraphy beginners,and formulated corresponding solutions to these problems,studied existing algorithms,and proposed improvements Opinion,and finally verify the improved algorithm proposed,so as to obtain a system that can make a reasonable evaluation of Lin's works.The main contents of this article are as follows:(1)An improved algorithm based on ZS fast parallel algorithm is proposed.This algorithm improves the problem of broken and burr when extracting the skeleton.The extracted skeleton image can well show the "morphology" of the calligraphy characters and can be better used in Calculation of image skeleton similarity.(2)This paper proposes a two-layer retrieval algorithm based on global features and local features.When the data is processed,the feature vector is hashed and the processed binary data is weighted.This method can ensure the precision of the recall Reduce the retrieval time and achieve the purpose of timely response.(3)Combining with the primary examination standard of calligraphy,this paper proposes a judging system based on image features.The system calculates the similarity of corresponding features from the three aspects of overall structural features,stroke shape features and font layout features,and uses the obtained The similarity value is used to evaluate the calligraphy works.(4)Build a computer evaluation system,input the user's temporary post works,compare with the calligrapher's works in the system,display the scores and comments on the user interface,and combine the comments to intuitively understand the shortcomings of the temporary post works,helping beginners to improve learning.
Keywords/Search Tags:skeleton extraction, double-layer retrieval algorithm, similarity calculation, computer evaluation system
PDF Full Text Request
Related items