Hierarchical lung field segmentation with joint shape and appearance sparse learning

Yeqin Shao, Yaozong Gao, Yanrong Guo, Yonghong Shi, Xin Yang, Dinggang Shen

Research output: Contribution to journalArticle

31 Citations (Scopus)

Abstract

Lung field segmentation in the posterior-anterior (PA) chest radiograph is important for pulmonary disease diagnosis and hemodialysis treatment. Due to high shape variation and boundary ambiguity, accurate lung field segmentation from chest radiograph is still a challenging task. To tackle these challenges, we propose a joint shape and appearance sparse learning method for robust and accurate lung field segmentation. The main contributions of this paper are: 1) a robust shape initialization method is designed to achieve an initial shape that is close to the lung boundary under segmentation; 2) a set of local sparse shape composition models are built based on local lung shape segments to overcome the high shape variations; 3) a set of local appearance models are similarly adopted by using sparse representation to capture the appearance characteristics in local lung boundary segments, thus effectively dealing with the lung boundary ambiguity; 4) a hierarchical deformable segmentation framework is proposed to integrate the scale-dependent shape and appearance information together for robust and accurate segmentation. Our method is evaluated on 247 PA chest radiographs in a public dataset. The experimental results show that the proposed local shape and appearance models outperform the conventional shape and appearance models. Compared with most of the state-of-the-art lung field segmentation methods under comparison, our method also shows a higher accuracy, which is comparable to the inter-observer annotation variation.

Original languageEnglish
Article number6737258
Pages (from-to)1761-1780
Number of pages20
JournalIEEE Transactions on Medical Imaging
Volume33
Issue number9
DOIs
Publication statusPublished - 2014 Jan 1

Fingerprint

Joints
Learning
Lung
Pulmonary diseases
Thorax
Observer Variation
Chemical analysis
Lung Diseases
Renal Dialysis

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Computer Science Applications
  • Radiological and Ultrasound Technology
  • Software

Cite this

Hierarchical lung field segmentation with joint shape and appearance sparse learning. / Shao, Yeqin; Gao, Yaozong; Guo, Yanrong; Shi, Yonghong; Yang, Xin; Shen, Dinggang.

In: IEEE Transactions on Medical Imaging, Vol. 33, No. 9, 6737258, 01.01.2014, p. 1761-1780.

Research output: Contribution to journalArticle

Shao, Yeqin ; Gao, Yaozong ; Guo, Yanrong ; Shi, Yonghong ; Yang, Xin ; Shen, Dinggang. / Hierarchical lung field segmentation with joint shape and appearance sparse learning. In: IEEE Transactions on Medical Imaging. 2014 ; Vol. 33, No. 9. pp. 1761-1780.
@article{2cc41310e2cd475ea5d177f8adb71caa,
title = "Hierarchical lung field segmentation with joint shape and appearance sparse learning",
abstract = "Lung field segmentation in the posterior-anterior (PA) chest radiograph is important for pulmonary disease diagnosis and hemodialysis treatment. Due to high shape variation and boundary ambiguity, accurate lung field segmentation from chest radiograph is still a challenging task. To tackle these challenges, we propose a joint shape and appearance sparse learning method for robust and accurate lung field segmentation. The main contributions of this paper are: 1) a robust shape initialization method is designed to achieve an initial shape that is close to the lung boundary under segmentation; 2) a set of local sparse shape composition models are built based on local lung shape segments to overcome the high shape variations; 3) a set of local appearance models are similarly adopted by using sparse representation to capture the appearance characteristics in local lung boundary segments, thus effectively dealing with the lung boundary ambiguity; 4) a hierarchical deformable segmentation framework is proposed to integrate the scale-dependent shape and appearance information together for robust and accurate segmentation. Our method is evaluated on 247 PA chest radiographs in a public dataset. The experimental results show that the proposed local shape and appearance models outperform the conventional shape and appearance models. Compared with most of the state-of-the-art lung field segmentation methods under comparison, our method also shows a higher accuracy, which is comparable to the inter-observer annotation variation.",
keywords = "Active shape model, chest radiograph, deformable segmentation, local appearance model, local sparse shape composition, sparse learning",
author = "Yeqin Shao and Yaozong Gao and Yanrong Guo and Yonghong Shi and Xin Yang and Dinggang Shen",
year = "2014",
month = "1",
day = "1",
doi = "10.1109/TMI.2014.2305691",
language = "English",
volume = "33",
pages = "1761--1780",
journal = "IEEE Transactions on Medical Imaging",
issn = "0278-0062",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "9",

}

TY - JOUR

T1 - Hierarchical lung field segmentation with joint shape and appearance sparse learning

AU - Shao, Yeqin

AU - Gao, Yaozong

AU - Guo, Yanrong

AU - Shi, Yonghong

AU - Yang, Xin

AU - Shen, Dinggang

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Lung field segmentation in the posterior-anterior (PA) chest radiograph is important for pulmonary disease diagnosis and hemodialysis treatment. Due to high shape variation and boundary ambiguity, accurate lung field segmentation from chest radiograph is still a challenging task. To tackle these challenges, we propose a joint shape and appearance sparse learning method for robust and accurate lung field segmentation. The main contributions of this paper are: 1) a robust shape initialization method is designed to achieve an initial shape that is close to the lung boundary under segmentation; 2) a set of local sparse shape composition models are built based on local lung shape segments to overcome the high shape variations; 3) a set of local appearance models are similarly adopted by using sparse representation to capture the appearance characteristics in local lung boundary segments, thus effectively dealing with the lung boundary ambiguity; 4) a hierarchical deformable segmentation framework is proposed to integrate the scale-dependent shape and appearance information together for robust and accurate segmentation. Our method is evaluated on 247 PA chest radiographs in a public dataset. The experimental results show that the proposed local shape and appearance models outperform the conventional shape and appearance models. Compared with most of the state-of-the-art lung field segmentation methods under comparison, our method also shows a higher accuracy, which is comparable to the inter-observer annotation variation.

AB - Lung field segmentation in the posterior-anterior (PA) chest radiograph is important for pulmonary disease diagnosis and hemodialysis treatment. Due to high shape variation and boundary ambiguity, accurate lung field segmentation from chest radiograph is still a challenging task. To tackle these challenges, we propose a joint shape and appearance sparse learning method for robust and accurate lung field segmentation. The main contributions of this paper are: 1) a robust shape initialization method is designed to achieve an initial shape that is close to the lung boundary under segmentation; 2) a set of local sparse shape composition models are built based on local lung shape segments to overcome the high shape variations; 3) a set of local appearance models are similarly adopted by using sparse representation to capture the appearance characteristics in local lung boundary segments, thus effectively dealing with the lung boundary ambiguity; 4) a hierarchical deformable segmentation framework is proposed to integrate the scale-dependent shape and appearance information together for robust and accurate segmentation. Our method is evaluated on 247 PA chest radiographs in a public dataset. The experimental results show that the proposed local shape and appearance models outperform the conventional shape and appearance models. Compared with most of the state-of-the-art lung field segmentation methods under comparison, our method also shows a higher accuracy, which is comparable to the inter-observer annotation variation.

KW - Active shape model

KW - chest radiograph

KW - deformable segmentation

KW - local appearance model

KW - local sparse shape composition

KW - sparse learning

UR - http://www.scopus.com/inward/record.url?scp=84906886315&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84906886315&partnerID=8YFLogxK

U2 - 10.1109/TMI.2014.2305691

DO - 10.1109/TMI.2014.2305691

M3 - Article

VL - 33

SP - 1761

EP - 1780

JO - IEEE Transactions on Medical Imaging

JF - IEEE Transactions on Medical Imaging

SN - 0278-0062

IS - 9

M1 - 6737258

ER -