Human action recognition using descriptor based on selective finite element analysis

Abstract

This paper proposes a novel local descriptor evaluated from the Finite Element Analysis for human action recognition. This local descriptor represents the distinctive human poses in the form of the stiffness matrix. This stiffness matrix gives the information of motion as well as shape change of the human body while performing an action. Initially, the human body is represented in the silhouette form. Most prominent points of the silhouette are then selected. This silhouette is discretized into several finite small triangle faces (elements) where the prominent points of the boundaries are the vertices of the triangles. The stiffness matrix of each triangle is then calculated. The feature vector representing the action video frame is constructed by combining all stiffness matrices of all possible triangles. These feature vectors are given to the Radial Basis Function-Support Vector Machine (RBF-SVM) classifier. The proposed method shows its superiority over other existing state-of-the-art methods on the challenging datasets Weizmann, KTH, Ballet, and IXMAS.

Modal Frequencies Based Human Action Recognition Using Silhouettes And Simplicial Elements

“…where D is displacement matrix subject to strain, t e is the thickness and S e is the area of the triangle and both are assumed constant. D is also called the strain displacement matrix which can be calculated by shape functions [21]. All the local stiffness matrices of triangle elements are united systematically to form the global stiffness matrix.…”

“…Modal frequencies of the human body silhouette reduce the feature dimensionality drastically when the reduced number of the mode of the frequency of the structure is selected. And thus, it takes less runtime as compared to literature [21]. 2.…”

“…The main disadvantage of these methodologies is that they give only motion information but no information about the structure. Methodologies reported in literature [15][16][17][18][19][20][21][22] based on silhouette analysis have a major contribution to human action recognition. Kapoor et al [15] have used average energy silhouette images whereas a hybrid classifier is used for action recognition by Mishra et al [16].…”

“…Even methodology that used hybrid descriptors having both global and local descriptors require fusion problems. A feature descriptor based on the stiffness matrix discussed in literature [21] is capable of representing both shape as well as motion features. The main limitation of this methodology was that it increases the feature dimensionality and also requires large memory space.…”

The process model of subjective quality assessment of videoconferencing in enterprise

“…Videos, the fastest-growing information carrier, will account for more than 80% of all Internet traffic by 2020 [2]. Video data has become easily accessible and dominant in the present time [3].…”