提出一种基于空时特征提取的人体步态识别算法。连续的特征子空间学习依次提取出步态的时间与空间特征：第一次特征子空间学习对步态的频域数据进行主成分分析，步态数据被转化为周期特征矢量；第二次特征子空间学习对步态数据的周期特征矢量形式进行主成分分析加线性判别分析的联合分析，步态数据被进一步转化为步态特征矢量。步态特征矢量同时包含运动的周期特征以及人体的形态特征，具有很强的识别能力。在USF步态数据库上的实验结果显示，该算法识别率较其他同类算法有明显提升。

A gait recognition method based on spatiotemporal feature extraction is proposed. Recursive subspace learning is used to extract both time and space feature of gait. In the first subspace learning, the periodic dynamic feature of gait is extracted by principal component analysis and sequence data is represented in the periodicity feature vector form. In the second subspace learning, principal component analysis plus linear discriminant analysis are applied to the oeriodicity feature vector representation of gait and sequence data is compressed into gait feature vector. gait feature vector is an effective representation because it contains both human dynamic and shape feature. Experimental result on the USF gait database shows that the proposed method achieves highly competitive performance with respect to other published gait recognition approaches.