Search Articles

Home / Articles

A novel approach for Human Action Recognition

. Dr. Upasana Sinha, Dileshwar Patel and Jyoti Yadav


Abstract

We propose a new neighborhood descriptor based on Finite Element Analysis for human development. That's the closest thing to a descriptor when it comes to human postures. If you're interested in learning more about this framework, please contact us at [email protected] or [email protected]. As soon as the schematic structure begins, it addresses the human constitution. The diagram's most unusual components are then selected. In order to detect the cutoff points, the vertices of the triangles are used. Each triangle's strength system is then determined. On the development video body trademark vector, all strength systems of each and every functional triangle are joined together as a reference. The RBF-SVM classifier uses these trademark vectors. Weizmann, KTH, Ballet and IXMAS datasets indicate that the suggested strategy outperforms other current methods.

 

Keywords:  finite element analysis (FEA), stiffness matrix, discretization, support vector machine

Download :