Efficient Articulated Trajectory Reconstruction Using Dynamic Programming and Filters - Robotics Institute Carnegie Mellon University

Efficient Articulated Trajectory Reconstruction Using Dynamic Programming and Filters

Jack L. Valmadre, Yingying Zhu, Sridha Sridharan, and Simon Lucey
Conference Paper, Proceedings of (ECCV) European Conference on Computer Vision, pp. 72 - 85, October, 2012

Abstract

This paper considers the problem of reconstructing the mo- tion of a 3D articulated tree from 2D point correspondences subject to some temporal prior. Hitherto, smooth motion has been encouraged using a trajectory basis, yielding a hard combinatorial problem with time complexity growing exponentially in the number of frames. Branch and bound strategies have previously attempted to curb this complexity whilst maintaining global optimality. However, they provide no guaran- tee of being more efficient than exhaustive search. Inspired by recent work which reconstructs general trajectories using compact high-pass filters, we develop a dynamic programming approach which scales lin- early in the number of frames, leveraging the intrinsically local nature of filter interactions. Extension to affine projection enables reconstruction without estimating cameras.

BibTeX

@conference{Valmadre-2012-17100,
author = {Jack L. Valmadre and Yingying Zhu and Sridha Sridharan and Simon Lucey},
title = {Efficient Articulated Trajectory Reconstruction Using Dynamic Programming and Filters},
booktitle = {Proceedings of (ECCV) European Conference on Computer Vision},
year = {2012},
month = {October},
pages = {72 - 85},
}