A comparative study of gesture recognition between RGB and HSV colors using time-space continuous dynamic programming

Abstract

This paper presents a comparative analysis of the color models. It used when identifying human skin in a gesture recognition system. The recognition system was based on time-space continuous dynamic programming (TSCDP). TSCDP performs optimal matching between point trajectories and video images. A point trajectory is a sequential pattern of pixels that can also be a representation of a type of gesture. Human skin can be represented either by RGB-color-based or HSV-color-based imagery. The score for gesture recognition depends on the selection of features used to represent the pixels in the point trajectory. We conducted experiments that compared the two candidate color systems, concluding that HSV-based skin features provide better scores than RGB-based skin features.