变形变换及其在手写体汉字整形中的应用

A Novel Deformation Transformation and Its Application to Handwritten Chinese Character Shape Correction

提出了一种新的一维变形变换方法并将之应用于手写体汉字的整形变换 .尽管不同书写风格的同类手写体汉字存在较大的随意性 ,但它们的拓扑结构是一致的 ,因此可将手写体汉字看成是一种可变形物体 .通过选择适当的变形参数 ,该变形变换可以将一个汉字变形为 2 4种不同的书写风格 ,利用一个变形系数可控制各种风格的变形程度 .该变形变换可以矫正手写体汉字中偏左、偏右、偏上、偏下、偏胖、偏瘦等各种位置变形 ,从而达到对手写体样本进行非线性规范化处理的效果 .实验结果表明 ,手写体汉字经变形变换后 ,识别率可提高大约 1.5 %左右 。

In this paper, a novel 1D deformation transformation for handwritten Chinese character shape correction is proposed. From our point of view, handwritten Chinese character can be regarded as a kind of deformable object. Although there are large variations in the same category of character for different handwriting styles, the basic topological structures of them are the same. With a suitable selection of deformation parameters, the 1D deformation transformation could deform a given handwritten Chinese character into 24 different handwriting styles. A deforming parameter controls the deformation degree for each style. The proposed deformation transformation could be applied as a non linear shape normalization method for Chinese character recognition to correct various position deformations such as left slanting, right slanting, top slanting, bottom slanting, too fat or too thin, etc in a handwritten Chinese character. Our experiment shows that the recognition rate can be improved by 1 5% or so by using the deformation transformation, demonstrating the effectiveness of the proposed transformation.