In this paper, we tackle the problem of action recognition using body skeletons extracted from video sequences. Our approach lies in the continuity of recent works representing video frames by Gramian matrices that describe a trajectory on the Riemannian manifold of positive-semidefinite matrices of fixed rank. In comparison with previous works, the manifold of fixed-rank positive-semidefinite matrices is here endowed with a different metric, and we resort to different algorithms for the curve fitting and temporal alignment steps. We evaluated our approach on three publicly available datasets (UTKinect-Action3D, KTH-Action and UAV-Gesture). The results of the proposed approach are competitive with respect to state-of-the-art methods, while only involving body skeletons.
@article{arxiv.1908.00646,
title = {Fitting, Comparison, and Alignment of Trajectories on Positive Semi-Definite Matrices with Application to Action Recognition},
author = {Benjamin Szczapa and Mohamed Daoudi and Stefano Berretti and Alberto Del Bimbo and Pietro Pala and Estelle Massart},
journal= {arXiv preprint arXiv:1908.00646},
year = {2019}
}
Comments
Updated version of the paper published in the workshop HBU2019. The differences with the published version are a few small corrections, mainly misleading notations for the distance function on p. 4, and missing square root in the expression for "d", in the Thm. on p. 4. Noticeable changes w. r. t. v1 and v2 on arxiv, please use this version instead