Research On The Chinese Off-line Handwriting Identification Based On The Characteristics Of Data Fusion

Handwriting identification is a technique that aims to decide the identity of writers according to the handwriting styles. With the improvement of applied field, handwriting identification becomes a very active area of the computer vision and pattern-recognition.In this paper, a novel valid method based on fusion technology and Support Vector Machine for writer identification is discussed. The research can offer technical support for the computer handwriting identification system, and expand further the scope of the application of handwriting identification and the field. The main contents of this thesis are as follows:The image processing of the handwriting is considered. A series of the processing algorithms are proposed in this paper. Its main idea consists of the image graying, the binary image and the image standardize. Further more, the experimental results show that these algorithms can achieve the better processing results.In this paper, we propose the shape feature method according to constructing the image feature vector which is normalized and spliced .The main methods of the shape feature Contains center features, eccentricity, solidity and extent. The center features reflect the relative of the writer's handwriting about the horizontal offset, and the other features show the writer's handwriting in the overlapping part of the stroke. The texture analysis is a hot approach in the automatic handwriting-based writer identification, so we introduce the basic knowledge and the common methods of the texture analysis. Further, we presented the detailed algorithm using the gray level co-occurrence matrix and use more efficient algorithm Gabor filter, and how to use this algorithm extracting the texture features. The algorithm is not only independent the text, but also unnecessary to segment the handwriting text. This advantage is consistent with the habit of handwriting identification.Based on the extraction of the shape features and the texture features, we normalized the features and deal with it though the technology of features fusion. Then, the Support Vector Machine is used to classify the handwriting features. We only discuss the Support Vector Machine in the linear situation and the nonlinear situation in this paper. In order to get a good ability of adaptability and resolution, we choose kernel function reasonably and set the slack variables and the penalty factors effectively. Further more, we can strict the maximize interval of the class in the classification. Last, the experimental results show that those algorithms achieved better processing results.