GRIDDS - a gait recognition image and depth dataset

Abstract

Several approaches based on human gait have been proposed in the literature, either for medical research reasons, smart surveillance, human-machine interaction, or other purposes, whose validation highly depends on the access to common input data through available datasets, enabling a coherent performance comparison. The advent of depth sensors leveraged the emergence of novel approaches and, consequently, the usage of new datasets. In this work we present the GRIDDS - A Gait Recognition Image and Depth Dataset, a new and publicly available gait depth-based dataset that can be used mostly for person and gender recognition purposes.