摘要
在步态识别过程中影响步态识别性能的因素很多。为了提高步态识别率,在最大差异伸展(MVU)算法的基础上,提出了一种监督MVU算法并应用于步态图像识别中。该方法能够通过线性变换找到一个最佳子空间,使不同子流形数据更分散、同一流形数据更紧密。由真实的步态图像数据库上的实验结果证实了所提出算法的有效性和可行性。
There are a number of covariate factors that affect recognition performance.In order to improve the gait recognition rate,based on maximum variance unfolding algorithm,this paper proposed a supervised MVU method and applied it to gait recognition.The proposed method could seek an optimal subspace where samples in different submanifolds were located further and samples in the same submanifolds were clustered closer.The experimental results on real-world gait image databases show the effectiveness and feasible of the proposed method.
出处
《计算机应用研究》
CSCD
北大核心
2012年第11期4390-4393,共4页
Application Research of Computers
基金
陕西省科技厅自然科学基金资助项目(2011JM8011)
关键词
步态识别
维数约简
最大差异伸展算法
监督最大差异伸展算法
gait recognition
dimensionality reduction
maximum variance unfolding(MVU)method
supervised maximum variance unfolding(SMVU)method