论训练样本集结构和稀疏表示分类算法的关系

近年来,基于表示法的人脸识别技术主要都集中在约束条件和字典学习.很少有研究用样本数据特征来确定基于表示分类算法的性能.本文定义了结构离散度,表示样本集的结构特征.实验结果表明,具有较高的结构离散度的集合能让一个分类算法获得更高的识别率.

Relationship between the representation-based classification algorithm and structure of the training sample set

In recent years,representation-based face-recognition techniques are focus mainly on constraint conditions and dictionary learning. Few researchers study which sample data features determine the performance of representation-based classification algorithms.we define the structure-scatter degree, which represents the structure features of training sample sets, said structure characteristics of sample set. Experimental results show that sets with a higher structure scatter more likely allows a classification algorithm to obtain a higher recognition rate.