特征提取是手写体汉字识别的关键，目前四方向网格特征已被实验证实是一种较好的手写体汉字特征。

针对通常的纵横弹性网格对汉字“撇、捺”笔画特征提取的不足．提出一种新的网格构造技术——对角弹性网格，它由45度和135度的对角直线构成，将汉字图像划分为多个菱形，能够很好地适应汉字在“撇、捺”方向的变化。将这两种网格单独，以及相互组合成双网格等情况分别进行手写体识别实验，实验结果验证了对角弹性网格的有效性和双弹性网格的高识别率性。

-Feature extraction is the key to handwritten Chinese character recognition. At present, the four-way grid feature has been proved to be a good handwritten Chinese character.

In view of the shortcomings of the traditional vertical and horizontal elastic grid on the extraction of Chinese characters　 write and tear . A new grid construction technique, diagonal elastic grid, is proposed, which consists of 45 degree and 135 degree diagonal straight lines. The Chinese characters are divided into multiple diamonds, which can adapt well to Chinese characters.　The direction of change. The experimental results verify the effectiveness of the diagonal elastic grids and the high recognition rate of the double elastic grids. The experimental results verify the effectiveness of the diagonal elastic grids and the high recognition rate of the birefringence grids.