Interactive prostate segmentation based on adaptive feature selection and manifold regularization

Sang Hyun Park, Yaozong Gao, Yinghuan Shi, Dinggang Shen

Research output: Contribution to journalArticle

Abstract

In this paper, we propose a new learning-based interactive editing method for prostate segmentation. Although many automatic methods have been proposed to segment the prostate, laborious manual correction is still required for many clinical applications due to the limited performance of automatic segmentation. The proposed method is able to flexibly correct wrong parts of the segmentation within a short time, even few scribbles or dots are provided. In order to obtain the robust correction with a few interactions, the discriminative features that can represent mid-level cues beyond image intensity or gradient are adaptively extracted from a local region of interest according to both the training set and the interaction. Then, the labeling problem is formulated as a semi-supervised learning task, which is aimed to preserve the manifold configuration between the labeled and unlabeled voxels. The proposed method is evaluated on a challenging prostate CT image data set with large shape and appearance variations. The automatic segmentation results originally with the average Dice of 0.766 were improved to the average Dice 0.866 after conducting totally 22 interactions for the 12 test images by using our proposed method.

Original languageEnglish
Pages (from-to)264-271
Number of pages8
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8679
Publication statusPublished - 2014 Jan 1
Externally publishedYes

Fingerprint

Supervised learning
Feature Selection
Labeling
Feature extraction
Regularization
Segmentation
Dice
Interaction
Interactive Methods
Semi-supervised Learning
CT Image
Voxel
Region of Interest
Gradient
Configuration

Keywords

  • Feature selection
  • Interactive segmentation
  • Manifold regularization
  • Prostate
  • Semisupervised learning

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

@article{0fd214563a1246b79ce4bc94c217ddde,
title = "Interactive prostate segmentation based on adaptive feature selection and manifold regularization",
abstract = "In this paper, we propose a new learning-based interactive editing method for prostate segmentation. Although many automatic methods have been proposed to segment the prostate, laborious manual correction is still required for many clinical applications due to the limited performance of automatic segmentation. The proposed method is able to flexibly correct wrong parts of the segmentation within a short time, even few scribbles or dots are provided. In order to obtain the robust correction with a few interactions, the discriminative features that can represent mid-level cues beyond image intensity or gradient are adaptively extracted from a local region of interest according to both the training set and the interaction. Then, the labeling problem is formulated as a semi-supervised learning task, which is aimed to preserve the manifold configuration between the labeled and unlabeled voxels. The proposed method is evaluated on a challenging prostate CT image data set with large shape and appearance variations. The automatic segmentation results originally with the average Dice of 0.766 were improved to the average Dice 0.866 after conducting totally 22 interactions for the 12 test images by using our proposed method.",
keywords = "Feature selection, Interactive segmentation, Manifold regularization, Prostate, Semisupervised learning",
author = "Park, {Sang Hyun} and Yaozong Gao and Yinghuan Shi and Dinggang Shen",
year = "2014",
month = "1",
day = "1",
language = "English",
volume = "8679",
pages = "264--271",
journal = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
issn = "0302-9743",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - Interactive prostate segmentation based on adaptive feature selection and manifold regularization

AU - Park, Sang Hyun

AU - Gao, Yaozong

AU - Shi, Yinghuan

AU - Shen, Dinggang

PY - 2014/1/1

Y1 - 2014/1/1

N2 - In this paper, we propose a new learning-based interactive editing method for prostate segmentation. Although many automatic methods have been proposed to segment the prostate, laborious manual correction is still required for many clinical applications due to the limited performance of automatic segmentation. The proposed method is able to flexibly correct wrong parts of the segmentation within a short time, even few scribbles or dots are provided. In order to obtain the robust correction with a few interactions, the discriminative features that can represent mid-level cues beyond image intensity or gradient are adaptively extracted from a local region of interest according to both the training set and the interaction. Then, the labeling problem is formulated as a semi-supervised learning task, which is aimed to preserve the manifold configuration between the labeled and unlabeled voxels. The proposed method is evaluated on a challenging prostate CT image data set with large shape and appearance variations. The automatic segmentation results originally with the average Dice of 0.766 were improved to the average Dice 0.866 after conducting totally 22 interactions for the 12 test images by using our proposed method.

AB - In this paper, we propose a new learning-based interactive editing method for prostate segmentation. Although many automatic methods have been proposed to segment the prostate, laborious manual correction is still required for many clinical applications due to the limited performance of automatic segmentation. The proposed method is able to flexibly correct wrong parts of the segmentation within a short time, even few scribbles or dots are provided. In order to obtain the robust correction with a few interactions, the discriminative features that can represent mid-level cues beyond image intensity or gradient are adaptively extracted from a local region of interest according to both the training set and the interaction. Then, the labeling problem is formulated as a semi-supervised learning task, which is aimed to preserve the manifold configuration between the labeled and unlabeled voxels. The proposed method is evaluated on a challenging prostate CT image data set with large shape and appearance variations. The automatic segmentation results originally with the average Dice of 0.766 were improved to the average Dice 0.866 after conducting totally 22 interactions for the 12 test images by using our proposed method.

KW - Feature selection

KW - Interactive segmentation

KW - Manifold regularization

KW - Prostate

KW - Semisupervised learning

UR - http://www.scopus.com/inward/record.url?scp=84921759769&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84921759769&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:84921759769

VL - 8679

SP - 264

EP - 271

JO - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

JF - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SN - 0302-9743

ER -