Study On Handwriting Identification Techniques Based On Texture Analysis

Writer Identification (WI) is a discipline which aims to decide the identity of writers according to the handwriting styles. The computerization of WI is the most important step to relax the heavy burden of document examiners and to achieve the goal of prevalent applications. The research of computer WI has a history of near 40 years, it is long in time and achievements in which is relatively rich. In recent years, social backgrounds urge more achievements in computer WI, and advances in correlated disciplines such as pattern recognition and artificial intelligence supply chances for the development of computer WI. Under this background, this thesis of master's degree studies and proposes a texture analysis method for handwriting image based on integer-to-integer wavelet transform by synthesizing methods that were excogitated. The main contents of this thesis are as follows.Firstly identity identification based on biologic characters (Biometric identification technology) is introduced, and 12 biologic statistical characters are compared. The application background and history of development of WI techniques are introduced in succession. The nature of the WI problem is analyzed, and then the strategy of building computer WI system is proposed. It is pointed out that an applicable computer WI system is human-computer interactive, composed of rough classification by computer and final decision by human experts.Texture analysis is applied in the manipulation, analysis and recognition of images widely. The thesis summarizes many methods that were excogitated in these several decades, in which wavelet transform has favorable time-frequency localization, scale transform and direction characteristic, and it has been a convincing tool for texture analysis as it can adapt mortal seeing system much easily.On character extraction, the thesis regards handwriting images as texture images, and excogitates analyzing handwriting images with (2,10) integer-to-integer wavelet transform innovatively, that is, analyzing textureimages with (2,10) integer-to-integer wavelet transform mainly, and denoting eigenvectors of handwriting images with integer-to-integer wavelet coefficients, the thesis defines an 18-dimension eigenvector.(2,10) integer-to-integer wavelet transform can extract the high pass of images effectively, high pass is the details of images, and analysis of texture images is mainly for the details. Using WT coefficients as the characters of images can do multi-resolution analysis, to reflect the details of texture better and to gain the ends of relaxing the heavy burden of document examiners. Experiments were made using Chinese handwriting from 27 different people, the contents are the same, and everyone has 2 to 4 shares, the characters are in the contexts instead of writing the same character, so they can represent handwriting in everyday life. In the environment of matlab6.5, selecting different representative characters to do the experiment gains preferable effect, depending on the text, the identification ratio can reach 98.60%.