中文摘要：

根据耦合度量学习方法能够直接处理不同集合的数据这一特性,将其应用到数据融合领域,提出了一种基于耦合度量学习的特征级融合方法.该方法首先通过增加对原始单个集合中具有相关关系的数据的优化处理,将耦合度量学习方法的目标函数改进成在耦合空间中所有具有相关关系的投影特征均彼此接近,从而使得这些特征的整体分布更满足特征级融合的需求,而后采用串行方式对特征进行融合,最终得到更加有效的特征用来分类识别.将上述方法应用到步态识别中,以解决步态识别中的数据融合问题.采用CASIA（B）步态数据库进行实验分析,结果表明该方法识别效果较好.

英文摘要：

The coupled metric learning method is applied to the data fusion field since it can directly deal w ith different datasets.A feature level fusion method based on the coupled metric learning is proposed.First,by adding the optimization of the correlation data in original single set,the objective function of the coupled metric learning method improves as all the projection features w ith correlation in the coupled space are close to each other.The overall distribution of these features becomes more suitable for feature level fusion.Then,the features are fused in serial mode.Finally,more effective features are obtained for classification.The proposed method is applied to solve the data fusion problem in gait recognition.The experiments and analysis are made based on the CASIA（B） gait database.The experimental results show that the proposed method can achieve good recognition results.