一种基于纹理特征的笔迹鉴别方法

Handwriting Identification Based on texture character

介绍了一种基于纹理特征的笔迹识别方法。把手写笔迹图像作为一种纹理看待,从而将笔迹识别问题转化为纹理识别问题。笔迹图像预处理采用基于像素点的切割方法,通过实验证明该方法可以较为完整的保存纹理信息。对于预处理后的纹理图像,再利用纹理处理技术和数学方法提取笔迹纹理特征,设计Gabor滤波器对特征图像鉴别,所有过程均使用MATLAB编程仿真实现。试验结果表明采用纹理识别的方法简单,避免了由于文字数量、纸张材质、字体颜色等因素的影响,显著提高了识别的准确度和笔迹识别应用的广度。

In this paper, a method of handwriting identification based on texture character is introduced. The handwriting image is considered as one of textures that achieves the change of handwriting identification and texture identification. Cutting method based on image element is used to handwriting image, which is justified by experiment. After the texture image is pretreated, the image texture processing technology and mathematical methods are used to distill the useful special texture, the paper designs Gabor filter to distinguish image which is achieved through the emulation of MATLAB. The testing results indicate that this method is more simple, which avoids the influence of the number of characters, the material of paper and the color of characters, it can greatly increase the identification of veracity and the range of the application.