Active Shape Model for Contour Detection of Anatomical Structure

Huiqi Li*, Qing Nie

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

Abstract

This chapter introduces an active shape model (ASM) for contour detection of anatomical structures. It presents the review of deformable models and ASM. The 2D deformable models can be categorized into two classes: free-form models and parametrical models. Free-form models are only constrained by some general local constraints. As there is no global structure, the models can represent arbitrary shapes. Parametrically deformable models are able to describe a specific characteristic shape by a set of parameters. The chapter describes the method of ASM. ASM is a parametric deformable model that a combines point distribution model (PDM) and an iterative refinement procedure. The chapter investigates two applications of ASM: the boundary detection of optic disk in fundus images and lens structure detection in lens images. Experiments show that the ASM is suitable for the boundary detection of anatomical structures.

Original languageEnglish
Title of host publicationBiomedical Image Understanding
Subtitle of host publicationMethods and Applications
Publisherwiley
Pages429-461
Number of pages33
ISBN (Electronic)9781118715321
ISBN (Print)9781118715154
DOIs
Publication statusPublished - 13 Feb 2015

Keywords

  • 2D deformable models
  • Active shape model (ASM)
  • Anatomical structure
  • Contour detection
  • Point distribution model (PDM)

Fingerprint

Dive into the research topics of 'Active Shape Model for Contour Detection of Anatomical Structure'. Together they form a unique fingerprint.

Cite this