Dense optical flow from multiple sparse candidate flows using two pass dynamic programming

T Smith, DW Redmill, CN Canagarajah, DR Bull

Research output: Chapter in Book/Report/Conference proceedingConference Contribution (Conference Proceeding)

Abstract

Optical flow forms an important initial processing stage for many machine vision tasks. A framework is presented for the recovery of dense optical flows from image sequences containing large motions. Sparse feature correspondences are used to assign multiple candidate optical flows to each image pixel. This set of flows is then augmented with additional perturbed flows to allow for non-rigid motions. An energy functional comprising of a matching term and smoothness term is then minimized using a two pass dynamic programming algorithm to produce a final smooth optical flow field. The proposed algorithm shows a clear increase in recovered optical flow accuracy when compared to a hierarchical approach and a brute force block matching approach of similar computational complexity.
Translated title of the contributionDense optical flow from multiple sparse candidate flows using two pass dynamic programming
Original languageEnglish
Title of host publicationInternational Conference on Visual Information Engineering, Xian, China
Place of PublicationXi'an,China
PublisherInstitution of Engineering and Technology (IET)
Pages203 - 208
ISBN (Print)9780863419140
DOIs
Publication statusPublished - Jul 2008
Event5th International Conference on Visual Information Engineering - Xian, China
Duration: 1 Jul 2008 → …

Conference

Conference5th International Conference on Visual Information Engineering
CountryChina
CityXian
Period1/07/08 → …

Bibliographical note

Conference Proceedings/Title of Journal: 5th International Conference on Visual Information Engineering, 2008 (VIE 2008)
Rose publication type: Conference contribution

Keywords

  • image motion analysis
  • feature extraction
  • machine vision

Fingerprint Dive into the research topics of 'Dense optical flow from multiple sparse candidate flows using two pass dynamic programming'. Together they form a unique fingerprint.

Cite this