Research Of Trace Handwriting Identification Technology Based On Texture Strengthening

Writer identification (WI) is a technology whitch aims to decide the identity ofwriters according to the handwriting styles.There is a wide range of theoretical basisand applications of WI technology. In addition to WI in specific areas such as publicsecurity professional WI, the application in the daily life has seen rapid development.For example, computer user login, credit card signature, e-commerce. From theperspective of identification, handwriting is a more stable behavior withcharacteristics of easy access (or non-contact), and acceptability. Therefore, researchgeneral WI system will have great significance.This article tries to trace the individual off-line handwriting (referring tosegment line or tiny stroke line of Chinese characters, characters, graphics or otherinformation) to identify the identity, this research can be applied to the computerapplication, mobile phone information,network login,credit card signature, socialwelfare recipients and other commercial environments. After analysis, this paperconsiders that the WI system has the following situations:1) a specific person with a specific handwriting;2)a particular person written in different handwriting;3)different people write specific handwritings;4)different handwriting written by different people.In this paper, for improving the handwriting image fidelity level for the next stepto prepare for extracting features and analysising,offline handwriting imagebinarization, denoising and normalization preprocessing, automatic layout analysishas been processed.In this paper,a small number of great significance of physical characteristics forWI has been selected by extracting and analysising by physical characteristics of thehandwriting image.In this paper,WI texture algorith has been analyzed and developed.As the amountof segment handwritings is limited, it is necessary to strengthen the texturecharacteristics.This paper based on the paper of many previous researchers which hasimproved the texture algorithm, and the image ranks of compression,image zoom,image superposition and other test has been tested.Finalized and implemented imageexpansion after image layout analysis algorithm to enhance the texture features.And achieve the desired effec.In this paper, the physical characteristics combine with texture feature todetermine the handwriting identification. After testing of the physical characteristics,this article uses only the physical characteristics of the overall tilt, eccentricity in thedecision integration of WI. Correctness of implementation of the algorithm of theselected physical characteristics and enhanced texture characteristics in the decisionfusion of WI will greatly improve the accept reate and lower the false accept rate.Thus the practical value of this system can be achieved.