Organ localization using joint AP/LAT view landmark consensus detection and Hierarchical active appearance models

Qi Song, Albert Montillo, Roshni Bhagalia, V. Srikrishnan

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Parsing 2D radiographs into anatomical regions is a challenging task with many applications. In the clinic, scans routinely include anterior-posterior (AP) and lateral (LAT) view radiographs. Since these orthogonal views provide complementary anatomic information, an integrated analysis can afford the greatest localization accuracy. To solve this integration we propose automatic landmark candidate detection, pruned by a learned geometric consensus detector model and refined by fitting a hierarchical active appearance organ model (H-AAM). Our main contribution is twofold. First, we propose a probabilistic joint consensus detection model which learns how landmarks in either or both views predict landmark locations in a given view. Second, we refine landmarks by fitting a joint H-AAM that learns how landmark arrangement and image appearance can help predict across views. This increases accuracy and robustness to anatomic variation. All steps require just seconds to compute and compared to processing the scouts separately, joint processing reduces mean landmark distance error from 27.3 mm to 15.7 mm in LAT view and from 12.7 mm to 11.2 mm in the AP view. The errors are comparable to human expert inter-observer variability and suitable for clinical applications such as personalized scan planning for dose reduction. We assess our method using a database of scout CT scans from 93 subjects with widely varying pathology.

Original languageEnglish (US)
Title of host publicationMedical Computer Vision
Subtitle of host publicationLarge Data in Medical Imaging - Third International MICCAI Workshop, MCV 2013, Revised Selected Papers
PublisherSpringer Verlag
Pages138-147
Number of pages10
Volume8331 LNCS
ISBN (Print)9783319055299
DOIs
StatePublished - Jan 1 2014
Event3rd International MICCAI Workshop on Medical Computer Vision, MCV 2013 - Nagoya, Japan
Duration: Sep 26 2013Sep 26 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8331 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other3rd International MICCAI Workshop on Medical Computer Vision, MCV 2013
CountryJapan
CityNagoya
Period9/26/139/26/13

Fingerprint

Active Appearance Models
Hierarchical Model
Landmarks
Lateral
Computerized tomography
Pathology
Processing
Predict
Parsing
Detectors
Planning
Observer
Dose
Arrangement
Detector
Robustness
Mm
Model

Keywords

  • Automatic landmark localization
  • CT
  • Hierarchical active appearance model
  • Image parsing
  • Organ localization
  • Rejection cascade

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Song, Q., Montillo, A., Bhagalia, R., & Srikrishnan, V. (2014). Organ localization using joint AP/LAT view landmark consensus detection and Hierarchical active appearance models. In Medical Computer Vision: Large Data in Medical Imaging - Third International MICCAI Workshop, MCV 2013, Revised Selected Papers (Vol. 8331 LNCS, pp. 138-147). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8331 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-05530-5_14

Organ localization using joint AP/LAT view landmark consensus detection and Hierarchical active appearance models. / Song, Qi; Montillo, Albert; Bhagalia, Roshni; Srikrishnan, V.

Medical Computer Vision: Large Data in Medical Imaging - Third International MICCAI Workshop, MCV 2013, Revised Selected Papers. Vol. 8331 LNCS Springer Verlag, 2014. p. 138-147 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8331 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Song, Q, Montillo, A, Bhagalia, R & Srikrishnan, V 2014, Organ localization using joint AP/LAT view landmark consensus detection and Hierarchical active appearance models. in Medical Computer Vision: Large Data in Medical Imaging - Third International MICCAI Workshop, MCV 2013, Revised Selected Papers. vol. 8331 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 8331 LNCS, Springer Verlag, pp. 138-147, 3rd International MICCAI Workshop on Medical Computer Vision, MCV 2013, Nagoya, Japan, 9/26/13. https://doi.org/10.1007/978-3-319-05530-5_14
Song Q, Montillo A, Bhagalia R, Srikrishnan V. Organ localization using joint AP/LAT view landmark consensus detection and Hierarchical active appearance models. In Medical Computer Vision: Large Data in Medical Imaging - Third International MICCAI Workshop, MCV 2013, Revised Selected Papers. Vol. 8331 LNCS. Springer Verlag. 2014. p. 138-147. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-05530-5_14
Song, Qi ; Montillo, Albert ; Bhagalia, Roshni ; Srikrishnan, V. / Organ localization using joint AP/LAT view landmark consensus detection and Hierarchical active appearance models. Medical Computer Vision: Large Data in Medical Imaging - Third International MICCAI Workshop, MCV 2013, Revised Selected Papers. Vol. 8331 LNCS Springer Verlag, 2014. pp. 138-147 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{c3ca219a77fa4add9cadea472e89a2b4,
title = "Organ localization using joint AP/LAT view landmark consensus detection and Hierarchical active appearance models",
abstract = "Parsing 2D radiographs into anatomical regions is a challenging task with many applications. In the clinic, scans routinely include anterior-posterior (AP) and lateral (LAT) view radiographs. Since these orthogonal views provide complementary anatomic information, an integrated analysis can afford the greatest localization accuracy. To solve this integration we propose automatic landmark candidate detection, pruned by a learned geometric consensus detector model and refined by fitting a hierarchical active appearance organ model (H-AAM). Our main contribution is twofold. First, we propose a probabilistic joint consensus detection model which learns how landmarks in either or both views predict landmark locations in a given view. Second, we refine landmarks by fitting a joint H-AAM that learns how landmark arrangement and image appearance can help predict across views. This increases accuracy and robustness to anatomic variation. All steps require just seconds to compute and compared to processing the scouts separately, joint processing reduces mean landmark distance error from 27.3 mm to 15.7 mm in LAT view and from 12.7 mm to 11.2 mm in the AP view. The errors are comparable to human expert inter-observer variability and suitable for clinical applications such as personalized scan planning for dose reduction. We assess our method using a database of scout CT scans from 93 subjects with widely varying pathology.",
keywords = "Automatic landmark localization, CT, Hierarchical active appearance model, Image parsing, Organ localization, Rejection cascade",
author = "Qi Song and Albert Montillo and Roshni Bhagalia and V. Srikrishnan",
year = "2014",
month = "1",
day = "1",
doi = "10.1007/978-3-319-05530-5_14",
language = "English (US)",
isbn = "9783319055299",
volume = "8331 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "138--147",
booktitle = "Medical Computer Vision",

}

TY - GEN

T1 - Organ localization using joint AP/LAT view landmark consensus detection and Hierarchical active appearance models

AU - Song, Qi

AU - Montillo, Albert

AU - Bhagalia, Roshni

AU - Srikrishnan, V.

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Parsing 2D radiographs into anatomical regions is a challenging task with many applications. In the clinic, scans routinely include anterior-posterior (AP) and lateral (LAT) view radiographs. Since these orthogonal views provide complementary anatomic information, an integrated analysis can afford the greatest localization accuracy. To solve this integration we propose automatic landmark candidate detection, pruned by a learned geometric consensus detector model and refined by fitting a hierarchical active appearance organ model (H-AAM). Our main contribution is twofold. First, we propose a probabilistic joint consensus detection model which learns how landmarks in either or both views predict landmark locations in a given view. Second, we refine landmarks by fitting a joint H-AAM that learns how landmark arrangement and image appearance can help predict across views. This increases accuracy and robustness to anatomic variation. All steps require just seconds to compute and compared to processing the scouts separately, joint processing reduces mean landmark distance error from 27.3 mm to 15.7 mm in LAT view and from 12.7 mm to 11.2 mm in the AP view. The errors are comparable to human expert inter-observer variability and suitable for clinical applications such as personalized scan planning for dose reduction. We assess our method using a database of scout CT scans from 93 subjects with widely varying pathology.

AB - Parsing 2D radiographs into anatomical regions is a challenging task with many applications. In the clinic, scans routinely include anterior-posterior (AP) and lateral (LAT) view radiographs. Since these orthogonal views provide complementary anatomic information, an integrated analysis can afford the greatest localization accuracy. To solve this integration we propose automatic landmark candidate detection, pruned by a learned geometric consensus detector model and refined by fitting a hierarchical active appearance organ model (H-AAM). Our main contribution is twofold. First, we propose a probabilistic joint consensus detection model which learns how landmarks in either or both views predict landmark locations in a given view. Second, we refine landmarks by fitting a joint H-AAM that learns how landmark arrangement and image appearance can help predict across views. This increases accuracy and robustness to anatomic variation. All steps require just seconds to compute and compared to processing the scouts separately, joint processing reduces mean landmark distance error from 27.3 mm to 15.7 mm in LAT view and from 12.7 mm to 11.2 mm in the AP view. The errors are comparable to human expert inter-observer variability and suitable for clinical applications such as personalized scan planning for dose reduction. We assess our method using a database of scout CT scans from 93 subjects with widely varying pathology.

KW - Automatic landmark localization

KW - CT

KW - Hierarchical active appearance model

KW - Image parsing

KW - Organ localization

KW - Rejection cascade

UR - http://www.scopus.com/inward/record.url?scp=84958528042&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84958528042&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-05530-5_14

DO - 10.1007/978-3-319-05530-5_14

M3 - Conference contribution

AN - SCOPUS:84958528042

SN - 9783319055299

VL - 8331 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 138

EP - 147

BT - Medical Computer Vision

PB - Springer Verlag

ER -