The study of pre-processing algorithm for improving performance of optical flow techniques on ultrasou nd image

Sung Min Kim, Ju Hwan Lee, Sung Yun Park, Seung Gyu Roh, Jae Hoon Jeong, Seung Ho Kang, Ho Chul Kang, Min Ji Seo

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In this study, we propose a novel pre-processing algorithm developed for improving the tracking efficiency of the optical flow method. The algorithm consists of median filtering, binarization, morphological processing, canny edge detection, contour detection and approximation. In order to evaluate whether our preprocessing algorithm increases the optical flow tracking capacity, this study applied the pre-processing algorithm to the Lucas-Kanade (LK) optical flow algorithm, and comparatively analyzed its images and tracking results with those of optical flow without pre-processing. It was observed that with pre-processing, the tracking performance derived from the LK optical flow algorithm showed better tracking accuracy. Such improvements seem to have resulted from the successful segmentation for characteristic areas and subdivision into inner and outer contour lines.

Original languageEnglish
Title of host publicationProceedings of the IASTED International Conference on Computational Bioscience, CompBio 2010
Pages640-646
Number of pages7
DOIs
StatePublished - 2010
EventIASTED International Conference on Computational Bioscience, CompBio 2010 - Cambridge, MA, United States
Duration: 1 Nov 20103 Nov 2010

Publication series

NameProceedings of the IASTED International Conference on Computational Bioscience, CompBio 2010

Conference

ConferenceIASTED International Conference on Computational Bioscience, CompBio 2010
Country/TerritoryUnited States
CityCambridge, MA
Period1/11/103/11/10

Keywords

  • Lucas-kanade
  • Optical flow
  • Pre-processing
  • Ultrasound image

Fingerprint

Dive into the research topics of 'The study of pre-processing algorithm for improving performance of optical flow techniques on ultrasou nd image'. Together they form a unique fingerprint.

Cite this