Header logo is

Robust dynamic motion estimation over time


Conference Paper


This paper presents a novel approach to incrementally estimating visual motion over a sequence of images. We start by formulating constraints on image motion to account for the possibility of multiple motions. This is achieved by exploiting the notions of weak continuity and robust statistics in the formulation of the minimization problem. The resulting objective function is non-convex. Traditional stochastic relaxation techniques for minimizing such functions prove inappropriate for the task. We present a highly parallel incremental stochastic minimization algorithm which has a number of advantages over previous approaches. The incremental nature of the scheme makes it truly dynamic and permits the detection of occlusion and disocclusion boundaries.

Award: (IEEE Computer Society Outstanding Paper Award)
Author(s): Black, M. J. and Anandan, P.
Book Title: Proc. Computer Vision and Pattern Recognition, CVPR-91,
Pages: 296-302
Year: 1991
Month: June

Department(s): Perceiving Systems
Bibtex Type: Conference Paper (inproceedings)
Paper Type: Conference

Address: Maui, Hawaii
Award Paper: IEEE Computer Society Outstanding Paper Award

Links: pdf


  title = {Robust dynamic motion estimation over time},
  author = {Black, M. J. and Anandan, P.},
  booktitle = {Proc. Computer Vision and Pattern Recognition, CVPR-91,},
  pages = {296-302},
  address = {Maui, Hawaii},
  month = jun,
  year = {1991},
  month_numeric = {6}