Gait-A Database: dataset of subjects walking

Abstract

Gait-A Database is a dataset of subjects walking over an 8 m hall. It contains both normal walking patterns along with some abnormal walking patterns feigned by the subjects. The dataset contains two folders, normal and abnormal, each containing the respective set of walking patterns. In each folder there are the different walking samples recorded. Each sample contains a silhouette folder containing the silhouette extracted (using Mean of Gaussians background subtraction technique) corresponding to each frame of the sequence.