Three-dimensional spatiotemporal features for fast content-based retrieval of focal liver lesions

Sharmili Roy, Yanling Chi, Jimin Liu, Sudhakar K Venkatesh, Michael S. Brown

Research output: Contribution to journalArticle

23 Citations (Scopus)

Abstract

Content-based image retrieval systems for 3-D medical datasets still largely rely on 2-D image-based features extracted from a few representative slices of the image stack. Most 2-D features that are currently used in the literature not only model a 3-D tumor incompletely but are also highly expensive in terms of computation time, especially for high-resolution datasets. Radiologist-specified semantic labels are sometimes used along with image-based 2-D features to improve the retrieval performance. Since radiological labels show large interuser variability, are often unstructured, and require user interaction, their use as lesion characterizing features is highly subjective, tedious, and slow. In this paper, we propose a 3-D image-based spatiotemporal feature extraction framework for fast content-based retrieval of focal liver lesions. All the features are computer generated and are extracted from four-phase abdominal CT images. Retrieval performance and query processing times for the proposed framework is evaluated on a database of 44 hepatic lesions comprising of five pathological types. Bull's eye percentage score above 85% is achieved for three out of the five lesion pathologies and for 98% of query lesions, at least one same type of lesion is ranked among the top two retrieved results. Experiments show that the proposed system's query processing is more than 20 times faster than other already published systems that use 2-D features. With fast computation time and high retrieval accuracy, the proposed system has the potential to be used as an assistant to radiologists for routine hepatic tumor diagnosis.

Original languageEnglish (US)
Article number6826549
Pages (from-to)2768-2778
Number of pages11
JournalIEEE Transactions on Biomedical Engineering
Volume61
Issue number11
DOIs
StatePublished - Nov 1 2014

Fingerprint

Content based retrieval
Query processing
Liver
Labels
Tumors
Image retrieval
Pathology
Feature extraction
Three-Dimensional Imaging
Semantics
Neoplasms
Databases
Experiments
Radiologists
Datasets

Keywords

  • 3-D spatiotemporal focal liver lesion representation
  • Clinical decision support system
  • contentbased image retrieval

ASJC Scopus subject areas

  • Biomedical Engineering
  • Medicine(all)

Cite this

Three-dimensional spatiotemporal features for fast content-based retrieval of focal liver lesions. / Roy, Sharmili; Chi, Yanling; Liu, Jimin; Venkatesh, Sudhakar K; Brown, Michael S.

In: IEEE Transactions on Biomedical Engineering, Vol. 61, No. 11, 6826549, 01.11.2014, p. 2768-2778.

Research output: Contribution to journalArticle

Roy, Sharmili ; Chi, Yanling ; Liu, Jimin ; Venkatesh, Sudhakar K ; Brown, Michael S. / Three-dimensional spatiotemporal features for fast content-based retrieval of focal liver lesions. In: IEEE Transactions on Biomedical Engineering. 2014 ; Vol. 61, No. 11. pp. 2768-2778.
@article{eb862da5e25e4d99885da6784139ad50,
title = "Three-dimensional spatiotemporal features for fast content-based retrieval of focal liver lesions",
abstract = "Content-based image retrieval systems for 3-D medical datasets still largely rely on 2-D image-based features extracted from a few representative slices of the image stack. Most 2-D features that are currently used in the literature not only model a 3-D tumor incompletely but are also highly expensive in terms of computation time, especially for high-resolution datasets. Radiologist-specified semantic labels are sometimes used along with image-based 2-D features to improve the retrieval performance. Since radiological labels show large interuser variability, are often unstructured, and require user interaction, their use as lesion characterizing features is highly subjective, tedious, and slow. In this paper, we propose a 3-D image-based spatiotemporal feature extraction framework for fast content-based retrieval of focal liver lesions. All the features are computer generated and are extracted from four-phase abdominal CT images. Retrieval performance and query processing times for the proposed framework is evaluated on a database of 44 hepatic lesions comprising of five pathological types. Bull's eye percentage score above 85{\%} is achieved for three out of the five lesion pathologies and for 98{\%} of query lesions, at least one same type of lesion is ranked among the top two retrieved results. Experiments show that the proposed system's query processing is more than 20 times faster than other already published systems that use 2-D features. With fast computation time and high retrieval accuracy, the proposed system has the potential to be used as an assistant to radiologists for routine hepatic tumor diagnosis.",
keywords = "3-D spatiotemporal focal liver lesion representation, Clinical decision support system, contentbased image retrieval",
author = "Sharmili Roy and Yanling Chi and Jimin Liu and Venkatesh, {Sudhakar K} and Brown, {Michael S.}",
year = "2014",
month = "11",
day = "1",
doi = "10.1109/TBME.2014.2329057",
language = "English (US)",
volume = "61",
pages = "2768--2778",
journal = "IEEE Transactions on Biomedical Engineering",
issn = "0018-9294",
publisher = "IEEE Computer Society",
number = "11",

}

TY - JOUR

T1 - Three-dimensional spatiotemporal features for fast content-based retrieval of focal liver lesions

AU - Roy, Sharmili

AU - Chi, Yanling

AU - Liu, Jimin

AU - Venkatesh, Sudhakar K

AU - Brown, Michael S.

PY - 2014/11/1

Y1 - 2014/11/1

N2 - Content-based image retrieval systems for 3-D medical datasets still largely rely on 2-D image-based features extracted from a few representative slices of the image stack. Most 2-D features that are currently used in the literature not only model a 3-D tumor incompletely but are also highly expensive in terms of computation time, especially for high-resolution datasets. Radiologist-specified semantic labels are sometimes used along with image-based 2-D features to improve the retrieval performance. Since radiological labels show large interuser variability, are often unstructured, and require user interaction, their use as lesion characterizing features is highly subjective, tedious, and slow. In this paper, we propose a 3-D image-based spatiotemporal feature extraction framework for fast content-based retrieval of focal liver lesions. All the features are computer generated and are extracted from four-phase abdominal CT images. Retrieval performance and query processing times for the proposed framework is evaluated on a database of 44 hepatic lesions comprising of five pathological types. Bull's eye percentage score above 85% is achieved for three out of the five lesion pathologies and for 98% of query lesions, at least one same type of lesion is ranked among the top two retrieved results. Experiments show that the proposed system's query processing is more than 20 times faster than other already published systems that use 2-D features. With fast computation time and high retrieval accuracy, the proposed system has the potential to be used as an assistant to radiologists for routine hepatic tumor diagnosis.

AB - Content-based image retrieval systems for 3-D medical datasets still largely rely on 2-D image-based features extracted from a few representative slices of the image stack. Most 2-D features that are currently used in the literature not only model a 3-D tumor incompletely but are also highly expensive in terms of computation time, especially for high-resolution datasets. Radiologist-specified semantic labels are sometimes used along with image-based 2-D features to improve the retrieval performance. Since radiological labels show large interuser variability, are often unstructured, and require user interaction, their use as lesion characterizing features is highly subjective, tedious, and slow. In this paper, we propose a 3-D image-based spatiotemporal feature extraction framework for fast content-based retrieval of focal liver lesions. All the features are computer generated and are extracted from four-phase abdominal CT images. Retrieval performance and query processing times for the proposed framework is evaluated on a database of 44 hepatic lesions comprising of five pathological types. Bull's eye percentage score above 85% is achieved for three out of the five lesion pathologies and for 98% of query lesions, at least one same type of lesion is ranked among the top two retrieved results. Experiments show that the proposed system's query processing is more than 20 times faster than other already published systems that use 2-D features. With fast computation time and high retrieval accuracy, the proposed system has the potential to be used as an assistant to radiologists for routine hepatic tumor diagnosis.

KW - 3-D spatiotemporal focal liver lesion representation

KW - Clinical decision support system

KW - contentbased image retrieval

UR - http://www.scopus.com/inward/record.url?scp=84908093866&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84908093866&partnerID=8YFLogxK

U2 - 10.1109/TBME.2014.2329057

DO - 10.1109/TBME.2014.2329057

M3 - Article

C2 - 24919041

AN - SCOPUS:84908093866

VL - 61

SP - 2768

EP - 2778

JO - IEEE Transactions on Biomedical Engineering

JF - IEEE Transactions on Biomedical Engineering

SN - 0018-9294

IS - 11

M1 - 6826549

ER -