Integrated Segmentation and Depth Ordering of Motion Layers in Image Sequences

David Tweed, Andrew Calway

Research output: Contribution to journalArticle (Academic Journal)peer-review

Abstract

We describe a method to segment and depth order motion layers simultaneously in an image sequence. Previous approaches have tended to ignore the depth ordering issue or treat it as a post-processing operation. We argue here that motion estimation and segmentation are crucially dependent on depth order and hence that the latter should form an integral part of any layering scheme. Using an explicit model of boundary ownership allowing simultaneous assignment of motions to regions and extraction of depth order, the method fuses colour region segmentations with motion estimates obtained via block correlation. The motion estimates are then updated using a depth-dependent partial correlation. Experiments show the approach is effective.
Translated title of the contributionIntegrated Segmentation and Depth Ordering of Motion Layers in Image Sequences
Original languageEnglish
Pages (from-to)322-331
JournalBritish Machine Vision Conference
Publication statusPublished - 2000

Bibliographical note

ISBN: 1901725138
Publisher: BMVA
Name and Venue of Conference: British Machine Vision Conference
Other identifier: 1000475

Fingerprint Dive into the research topics of 'Integrated Segmentation and Depth Ordering of Motion Layers in Image Sequences'. Together they form a unique fingerprint.

Cite this