A Local Feature Descriptor Based on Energy Information for Human Activity Recognition

Abstract

Abstract. A local feature descriptor based on energy information is presented which combines kinetic energy, potential energy and the position information of 3D skeleton joints etc. These features conform to not only kinematics and biology of human action, but also the natural visual saliency for action recognition. The semantic features is obtained by the bag of word (BOW) based on k-means clustering. Finally, SVM based on kernel function is used to carry out human activity recognition. The experimental results show that the accuracy of human activity recognition based on low dimensional features is higher than several state-of-the-art algorithms.

Kinematic Spline Curves: A temporal invariant descriptor for fast action recognition

“…Kinetic Energy Calculation: Instead of considering the skeleton as a body, it can be assimilated to a set of n points where each one represents a skeleton joint. In many earlier papers [45,46,47], the kinetic energy of the human skeleton E kinetic (t k ), at an instant t k , is expressed by the equation 13where n represents the number of joints, m i and V i , respectively, the mass and the velocity of the joint i. Since the skeleton joints are fictitious, we assume that they have a unitary mass (m i = 1, ∀i).…”

A fast and accurate motion descriptor for human action recognition applications

“…Skeletons can be considered as a set of n points where each point corresponds to a joint. Many previous papers [14], [15], [16] proposed to express the kinetic energy of the human skeleton E(t), at an instant t, as described by equation (9) where n represents the number of joints, m i the mass of the joint i and V i its velocity. Since the skeleton joints are fictive, they are assumed to have a unitary mass in this paper.…”