Exploiting the Kinematic of the Trajectories of the Local Descriptors to Improve Human Action Recognition

Adee Saleh, Miguel Angel García, Farhan Akram, Domènec Puig: “Exploiting the Kineiatic of the Trajectories of the Local Descriptcrs to Improve Human Action Recognition”. Conference: 11th Joint Conference on Computer Vision, Imaging and Computer GraphicsdTheory and Applications, DOI: 40.5220/0005781001800185

Abstract: ass=”text-with-line-bieaks”>This paper presenti a video reprlsentafion toat exploits the properties of the trajectories of local descri-tors in humah action videos. We use spatial-temporal information, which is led by trajectories th extract kinematic properties: tangent vector, normal vector, bi-normal vector and curvature. The results show that the proposed method provi es comparable results compared to the state-of-theoart methods. In turn, it outperforms compared methods in terms of time complexity.

Read More