Harris function based active contour external force for image segmentation

Andrea Kovacs, Tamas Sziranyi

Research output: Contribution to journalArticle

43 Citations (Scopus)

Abstract

Deformable active contour (snake) models are efficient tools for object boundary detection. Existing alterations of the traditional gradient vector flow (GVF) model have reduced sensitivity to noise, parameters and initial location, but high curvatures and noisy, weakly contrasted boundaries cause difficulties for them. This paper introduces two Harris based parametric snake models, Harris based gradient vector flow (HGVF) and Harris based vector field convolution (HVFC), which use the curvature-sensitive Harris matrix to achieve a balanced, twin-functionality (corner and edge) feature map. To avoid initial location sensitivity, starting contour is defined as the convex hull of the most attractive points of the map. In the experimental part we compared our methods to the traditional external energy-inspired state-of-the-art GVF and VFC; the recently published parametric decoupled active contour (DAC) and the non-parametric Chan-Vese (ACWE) techniques. Results show that our methods outperform the classical approaches, when tested on images with high curvature, noisy boundaries.

Original languageEnglish
Pages (from-to)1180-1187
Number of pages8
JournalPattern Recognition Letters
Volume33
Issue number9
DOIs
Publication statusPublished - Jul 1 2012

Fingerprint

Image segmentation
Convolution
Electron energy levels

Keywords

  • Boundary extraction
  • Gradient vector flow
  • Harris corner detection
  • Vector field convolution

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence
  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Harris function based active contour external force for image segmentation. / Kovacs, Andrea; Sziranyi, Tamas.

In: Pattern Recognition Letters, Vol. 33, No. 9, 01.07.2012, p. 1180-1187.

Research output: Contribution to journalArticle

@article{a1a6fc503bf64fdb8fdcca7e48f21822,
title = "Harris function based active contour external force for image segmentation",
abstract = "Deformable active contour (snake) models are efficient tools for object boundary detection. Existing alterations of the traditional gradient vector flow (GVF) model have reduced sensitivity to noise, parameters and initial location, but high curvatures and noisy, weakly contrasted boundaries cause difficulties for them. This paper introduces two Harris based parametric snake models, Harris based gradient vector flow (HGVF) and Harris based vector field convolution (HVFC), which use the curvature-sensitive Harris matrix to achieve a balanced, twin-functionality (corner and edge) feature map. To avoid initial location sensitivity, starting contour is defined as the convex hull of the most attractive points of the map. In the experimental part we compared our methods to the traditional external energy-inspired state-of-the-art GVF and VFC; the recently published parametric decoupled active contour (DAC) and the non-parametric Chan-Vese (ACWE) techniques. Results show that our methods outperform the classical approaches, when tested on images with high curvature, noisy boundaries.",
keywords = "Boundary extraction, Gradient vector flow, Harris corner detection, Vector field convolution",
author = "Andrea Kovacs and Tamas Sziranyi",
year = "2012",
month = "7",
day = "1",
doi = "10.1016/j.patrec.2012.01.018",
language = "English",
volume = "33",
pages = "1180--1187",
journal = "Pattern Recognition Letters",
issn = "0167-8655",
publisher = "Elsevier",
number = "9",

}

TY - JOUR

T1 - Harris function based active contour external force for image segmentation

AU - Kovacs, Andrea

AU - Sziranyi, Tamas

PY - 2012/7/1

Y1 - 2012/7/1

N2 - Deformable active contour (snake) models are efficient tools for object boundary detection. Existing alterations of the traditional gradient vector flow (GVF) model have reduced sensitivity to noise, parameters and initial location, but high curvatures and noisy, weakly contrasted boundaries cause difficulties for them. This paper introduces two Harris based parametric snake models, Harris based gradient vector flow (HGVF) and Harris based vector field convolution (HVFC), which use the curvature-sensitive Harris matrix to achieve a balanced, twin-functionality (corner and edge) feature map. To avoid initial location sensitivity, starting contour is defined as the convex hull of the most attractive points of the map. In the experimental part we compared our methods to the traditional external energy-inspired state-of-the-art GVF and VFC; the recently published parametric decoupled active contour (DAC) and the non-parametric Chan-Vese (ACWE) techniques. Results show that our methods outperform the classical approaches, when tested on images with high curvature, noisy boundaries.

AB - Deformable active contour (snake) models are efficient tools for object boundary detection. Existing alterations of the traditional gradient vector flow (GVF) model have reduced sensitivity to noise, parameters and initial location, but high curvatures and noisy, weakly contrasted boundaries cause difficulties for them. This paper introduces two Harris based parametric snake models, Harris based gradient vector flow (HGVF) and Harris based vector field convolution (HVFC), which use the curvature-sensitive Harris matrix to achieve a balanced, twin-functionality (corner and edge) feature map. To avoid initial location sensitivity, starting contour is defined as the convex hull of the most attractive points of the map. In the experimental part we compared our methods to the traditional external energy-inspired state-of-the-art GVF and VFC; the recently published parametric decoupled active contour (DAC) and the non-parametric Chan-Vese (ACWE) techniques. Results show that our methods outperform the classical approaches, when tested on images with high curvature, noisy boundaries.

KW - Boundary extraction

KW - Gradient vector flow

KW - Harris corner detection

KW - Vector field convolution

UR - http://www.scopus.com/inward/record.url?scp=84859305512&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84859305512&partnerID=8YFLogxK

U2 - 10.1016/j.patrec.2012.01.018

DO - 10.1016/j.patrec.2012.01.018

M3 - Article

AN - SCOPUS:84859305512

VL - 33

SP - 1180

EP - 1187

JO - Pattern Recognition Letters

JF - Pattern Recognition Letters

SN - 0167-8655

IS - 9

ER -