Abstract. We suggest a variational method for the joint estimation of optic flow and the segmentation of the image into regions of similar motion. It makes use of the level set framework following the idea of motion competition, which is extended to non-parametric motion. Moreover, we automatically determine an appropriate initialization and the number of regions by means of recursive two-phase splits with higher order region models. The method is further extended to the spatiotemporal setting and the use of additional cues like the gray value or color for the segmentation. It need not fear a quantitative comparison to pure optic flow estimation techniques: For the popular Yosemite sequence with clouds we obtain the currently most accurate result. We further uncover a mistake in the ground truth. Coarsely correcting this, we get an average angular error below 1 degree. 1
Description
CiteSeerX — Variational motion segmentation with level sets
%0 Conference Paper
%1 Brox06variationalmotion
%A Brox, Thomas
%A Bruhn, Andrés
%A Weickert, Joachim
%B in ECCV
%D 2006
%I Springer
%K motion_analysis motion_segmentation
%P 471--483
%T Variational motion segmentation with level sets
%U http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.86.5468&rank=20
%X Abstract. We suggest a variational method for the joint estimation of optic flow and the segmentation of the image into regions of similar motion. It makes use of the level set framework following the idea of motion competition, which is extended to non-parametric motion. Moreover, we automatically determine an appropriate initialization and the number of regions by means of recursive two-phase splits with higher order region models. The method is further extended to the spatiotemporal setting and the use of additional cues like the gray value or color for the segmentation. It need not fear a quantitative comparison to pure optic flow estimation techniques: For the popular Yosemite sequence with clouds we obtain the currently most accurate result. We further uncover a mistake in the ground truth. Coarsely correcting this, we get an average angular error below 1 degree. 1
@inproceedings{Brox06variationalmotion,
abstract = {Abstract. We suggest a variational method for the joint estimation of optic flow and the segmentation of the image into regions of similar motion. It makes use of the level set framework following the idea of motion competition, which is extended to non-parametric motion. Moreover, we automatically determine an appropriate initialization and the number of regions by means of recursive two-phase splits with higher order region models. The method is further extended to the spatiotemporal setting and the use of additional cues like the gray value or color for the segmentation. It need not fear a quantitative comparison to pure optic flow estimation techniques: For the popular Yosemite sequence with clouds we obtain the currently most accurate result. We further uncover a mistake in the ground truth. Coarsely correcting this, we get an average angular error below 1 degree. 1},
added-at = {2014-07-29T14:14:30.000+0200},
author = {Brox, Thomas and Bruhn, Andrés and Weickert, Joachim},
biburl = {https://www.bibsonomy.org/bibtex/275126b49bee075bcd73073b11f5d1d98/alex_ruff},
booktitle = {in ECCV},
description = {CiteSeerX — Variational motion segmentation with level sets},
interhash = {3ea6d7aadb285f8cdf7917fe98dcf2ff},
intrahash = {75126b49bee075bcd73073b11f5d1d98},
keywords = {motion_analysis motion_segmentation},
pages = {471--483},
publisher = {Springer},
timestamp = {2014-07-29T14:14:30.000+0200},
title = {Variational motion segmentation with level sets},
url = {http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.86.5468&rank=20},
year = 2006
}