用时空能量图和小波变换方法表征和识别步态

该文提出一种基于时空能量图和小波变换的步态特征表达和步态识别方法，步态特征由时空能量图经过小波变换后来表示，不必进行严格的时间配准或步态周期的定位。时空能量图将一个序列的多帧图像用一幅灰度图替代，集成了人的运动信息中时间与空间变化的特点，减少了特征的数据量，经证明，时空能量图对噪声不敏感。实验结果表明该算法具有较好的识别性能和较低的空间需求和计算量。

Gait Representation and Recognition Using Spatio-Temporal Energy Image and Wavelet Transformation

Aiming at the dependency of the gait recognition algorithms on the time alignment and the gait cycle extraction, a novel gait representation is proposed. The spatio-temporal energy (STE) image and the wavelet transformation (WT) are used to deal with the problem of time-dependence and time-invariant covariation. The extracted features cover the physiological characteristics and the behavior characteristics. Recognition results demonstrate that the method has good performance, as well as the merit of low space and time requirement.