A new adaptive B-spline VFC snake for object contour extraction

Hoang Nam Nguyen*, An-Chen Lee

*Corresponding author for this work

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

Abstract

We propose a new adaptive B-spline VFC Snake model for object contour extraction. Bing Li et al. proposed vector field convolution (VFC) snake which has the advantages of superior noise robustness, reducing computational cost, and large capture range. However, it suffers from slow convergence speed due to large number of control points, as well as from difficulties in determining the weight factors associated to the internal energies constraining the smoothness of the curve. There is also no relevant criterion to determine the number of control points in VFC snake method. Our alternative approach expresses the curve as a non-uniform B-spline, in which fewer parameters are required and most importantly, internal energy calculation is eliminated because the smoothness is implicitly built into the model. A novel formulation of control points' movement estimation was established based on the least square fitting of non-uniform B-spline curve and VFC external force for the snake evolution process. A novel strategy of adding control points quickly matches the snake to desired complex shapes. Experimental results demonstrate the capability of adaptive shape description with high convergence speed of the proposed model.

Original languageEnglish
Title of host publicationComputational Collective Intelligence
Subtitle of host publicationTechnologies and Applications - Second International Conference, ICCCI 2010, Proceedings
Pages29-36
Number of pages8
EditionPART 3
DOIs
StatePublished - 3 Dec 2010
Event2nd International Conference on Computational Collective Intelligence - Technologies and Applications, ICCCI 2010 - Kaohsiung, Taiwan
Duration: 10 Nov 201012 Nov 2010

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 3
Volume6423 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2nd International Conference on Computational Collective Intelligence - Technologies and Applications, ICCCI 2010
CountryTaiwan
CityKaohsiung
Period10/11/1012/11/10

Keywords

  • Active contour
  • B-snake
  • B-spline
  • Computer vision
  • Contour extraction

Fingerprint Dive into the research topics of 'A new adaptive B-spline VFC snake for object contour extraction'. Together they form a unique fingerprint.

Cite this