Matching Kinematic Structures

Kinematic Structure Correspondences via Hypergraph Matching

Hyung Jin Chang     Tobias Fischer     Maxime Petit     Martina Zambelli     Yiannis Demiris
Imperial College London
KS_fig1
The proposed framework reliably builds up kinematic structure correspondence matches across heterogeneous objects captured with different sensors. Our method can for example find correspondences between a upper-body dancing human in a 2D grey image sequence, the iCub and NAO humanoid robots in 2D RGB videos, and a dancing human in depth image sequences.
KS_fig2
Various kinematic structure correspondence matching results using the proposed method

[Abstract]

In this paper, we present a novel framework for finding the kinematic structure correspondence between two objects in videos via hypergraph matching. In contrast to prior appearance and graph alignment based matching methods which have been applied among two similar static images, the proposed method finds correspondences between two dynamic kinematic structures of heterogeneous objects in videos. Our main contributions can be summarised as follows: (i) casting the kinematic structure correspondence problem into a hypergraph matching problem, incorporating multi-order similarities with normalising weights, (ii) structural topology similarity measure by a new topology constrained subgraph isomorphism aggregation, (iii) kinematic correlation measure between pairwise nodes, (iv) combinatorial local motion similarity measure using geodesic distance on the Riemannian manifold. We demonstrate the robustness and accuracy of our method through a number of experiments on synthetic and real data, showing that various other methods are out-performed.

[Paper]

PAPER | Supplementary


[Video]


[Dataset & Code]

Imperial-PRL KSC Dataset [Download]


[Bibtex]

@inproceedings{ChangCVPR2016KinematicStructure,
author = {Hyung Jin Chang and Tobias Fischer and Maxime Petit and Martina Zambelli and Yiannis Demiris},
title = {Kinematic Structure Correspondences via Hypergraph Matching},
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
pages = {4216–4425},
month = {June},
year = {2016}
}

[Acknowledgement]

This work was supported in part by the EU FP7 project WYSIWYD under Grant 612139.
We thank Dr. Minsu Cho for fruitful discussions.

For questions, contact hj.chang@imperial.ac.uk
Advertisements