一种基于流形学习的手写体数字识别

Handwritten Numeral Recognition Based on Manifold Learning

针对手写字符识别中由于书写习惯和风格不同造成字符模式不稳定的问题,提出了一种基于流形学习的手写体数字识别方法。算法首先利用局部线性嵌入(LLE)对手写体数字图像进行字符特征的降维,然后再对降维后的特征进行分类识别。通过对MINST库中手写体数字数据库上的实验结果表明,利用LLE降维后的特征能够有效地区分字符,识别率达到91.7%,并能够发现高维空间的低维嵌入流形。

To ovecome the instability of handwritten character model that is caused by different writing style, a novel approach based on manifold learning is proposed in this paper. First,Locally linear ernbedding（LLE） algorithm is used to reduce the dimensionality of input feature. Then the reduced feature is classified by simple classifier. Finally,the algorithm proposed in this paper is tested on the characters in MINST character database,and the experimental results demonstrate that the method can effectively improve the recognition rate to 91.7% of handwirtten digits and provide a new approach to the research for handwirtten digits recognition.