EigenGait: Motion-Based Recognition of People Using Image Self-Similarity

We present a novel technique for motion-based recognition of individual gaits in monocular sequences. Recent work has suggested that the image self-similarity plot of a moving person/object is a projection of its planar dynamics. Hence we expect that these plots encode much information about gait motion patterns, and that they can serve as good discriminants between different types of gait. We propose a method for gait recognition that uses similarity plots the same way that face images are used in eigenface-based face recognition techniques. Specifically, we first apply Principal Component Analysis (PCA) to a set of training similarity plots. This maps them to a lower dimensional space with less unwanted variation. A supervised clustering technique is then used to group these training samples into classes (clusters) in this space. Gait recognition is subsequently done by assigning the similarity plot of an input gait to the closest cluster. We test our method on a data set of 42 sequences at 30 fps each, depicting six different walking subjects. We use k-means clustering to build the classifier and the leave-one-out cross-validation technique to obtain an unbiased estimate of the recognition rate of about 93% (39