Gesture recognition using Integral Imaging
Filiberto Pla, P. Carmona, Eva Salvador-Balaguer, Bahram Javidi · 2017
In this paper we present some results summarizing our previous work during the last few years in the analysis of Synthetic Aperture Integral Imaging (SAII) performance for human gesture recognition, with a particular emphasis in its accuracy in the presence of occlusions, also when comparing it with sensors and strategies that are commonly used for these tasks, like the RGB-D sensors. Experiments show that its performance compares favorably against monocular imaging and to RGB-D sensors and that gesture classification results support the increase in performance for SAII particularly under partial occlusions.