基于矩阵SVD的手写数字分类及其特征区域的SVD优化

SVD Decomposition-based Approach for Handwritten Digits Classification

本文提出了三种利用矩阵的SVD分解来实现手写数字分类的方法。第一种方法基于矩阵的svd分解,将每一个数字用一个一维向量表示而对某一类的数字组成的矩阵进行奇异值分解(SVD);第二种方法是对特征区域的SVD优化,是对第一种方法的基础上进行的改进;两种方法最终的分类实现都是运用相应的最小二乘问题的方法求解。采用USPS的手写数字数据库进行训练和测试,并对两种方法的结果进行了简要对比。

This paper proposes two methods to realize using matrix handwritten digit classification methods. A method based on a matrix decomposition SVD, each with a number of one-dimensional vector and a matrix consisting of the number of classes for the singular value decomposition(SVD) means;, two methods to achieve the final classification is appropriate to use the least squares method for solving the problem. Using USPS handwritten digital database for training and testing, and the results of the two methods are briefly compared.