关键词: 手写数字图像 统计特征 核主成分分析 去噪

Abstract:

Handwritten digits de-noising is the principal preprocessing of handwritten digits recognition. In this paper we give the noise feature from the stati  stical point of view, de-noise the image using the kernel principal component analysis with the feature,and analyze the results of this method. Finally   we give the future research consideration of the handwritten digits de-noising problem.

Key words: handwritten digits, statistical feature, kernel principal component analysis;de-noising