Efficient small blob detection based on local convexity, intensity and shape information

Min Zhang, Teresa Wu, Scott C. Beeman, Luise Cullen-McEwen, John F. Bertram, Jennifer R. Charlton, Edwin Baldelomar, Kevin M. Bennett

Research output: Contribution to journalArticleResearchpeer-review

Abstract

The identification of small structures (blobs) from medical images to quantify clinically relevant features, such as size and shape, is important in many medical applications. One particular application explored here is the automated detection of kidney glomeruli after targeted contrast enhancement and magnetic resonance imaging. We propose a computationally efficient algorithm, termed the Hessian-based Difference of Gaussians (HDoG), to segment small blobs (e.g. glomeruli from kidney) from 3D medical images based on local convexity, intensity and shape information. The image is first smoothed and pre-segmented into small blob candidate regions based on local convexity. Two novel 3D regional features (regional blobness and regional flatness) are then extracted from the candidate regions. Together with regional intensity, the three features are used in an unsupervised learning algorithm for auto post-pruning. HDoG is first validated in a 2D form and compared with other three blob detectors from literature, which are generally for 2D images only. To test the detectability of blobs from 3D images, 240 sets of simulated images are rendered for scenarios mimicking the renal nephron distribution observed in contrast-enhanced, 3D MRI. The results show a satisfactory performance of HDoG in detecting large numbers of small blobs. Two sets of real kidney 3D MR images (6 rats, 3 human) are then used to validate the applicability of HDoG for glomeruli detection. By comparing MRI to stereological measurements, we verify that HDoG is a robust and efficient unsupervised technique for 3D blobs segmentation.
Original languageEnglish
Pages (from-to)1127-1137
Number of pages11
JournalIEEE Transactions on Medical Imaging
Volume35
Issue number4
DOIs
Publication statusPublished - Apr 2016

Keywords

  • kidney
  • machine learning
  • quantification and estimation
  • segmentation
  • shape analysis

Cite this

Zhang, Min ; Wu, Teresa ; Beeman, Scott C. ; Cullen-McEwen, Luise ; Bertram, John F. ; Charlton, Jennifer R. ; Baldelomar, Edwin ; Bennett, Kevin M. / Efficient small blob detection based on local convexity, intensity and shape information. In: IEEE Transactions on Medical Imaging. 2016 ; Vol. 35, No. 4. pp. 1127-1137.
@article{7df8c1b748bc43b49c37558dec392b18,
title = "Efficient small blob detection based on local convexity, intensity and shape information",
abstract = "The identification of small structures (blobs) from medical images to quantify clinically relevant features, such as size and shape, is important in many medical applications. One particular application explored here is the automated detection of kidney glomeruli after targeted contrast enhancement and magnetic resonance imaging. We propose a computationally efficient algorithm, termed the Hessian-based Difference of Gaussians (HDoG), to segment small blobs (e.g. glomeruli from kidney) from 3D medical images based on local convexity, intensity and shape information. The image is first smoothed and pre-segmented into small blob candidate regions based on local convexity. Two novel 3D regional features (regional blobness and regional flatness) are then extracted from the candidate regions. Together with regional intensity, the three features are used in an unsupervised learning algorithm for auto post-pruning. HDoG is first validated in a 2D form and compared with other three blob detectors from literature, which are generally for 2D images only. To test the detectability of blobs from 3D images, 240 sets of simulated images are rendered for scenarios mimicking the renal nephron distribution observed in contrast-enhanced, 3D MRI. The results show a satisfactory performance of HDoG in detecting large numbers of small blobs. Two sets of real kidney 3D MR images (6 rats, 3 human) are then used to validate the applicability of HDoG for glomeruli detection. By comparing MRI to stereological measurements, we verify that HDoG is a robust and efficient unsupervised technique for 3D blobs segmentation.",
keywords = "kidney, machine learning, quantification and estimation, segmentation, shape analysis",
author = "Min Zhang and Teresa Wu and Beeman, {Scott C.} and Luise Cullen-McEwen and Bertram, {John F.} and Charlton, {Jennifer R.} and Edwin Baldelomar and Bennett, {Kevin M.}",
year = "2016",
month = "4",
doi = "10.1109/TMI.2015.2509463",
language = "English",
volume = "35",
pages = "1127--1137",
journal = "IEEE Transactions on Medical Imaging",
issn = "0278-0062",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
number = "4",

}

Efficient small blob detection based on local convexity, intensity and shape information. / Zhang, Min; Wu, Teresa; Beeman, Scott C.; Cullen-McEwen, Luise; Bertram, John F.; Charlton, Jennifer R.; Baldelomar, Edwin; Bennett, Kevin M.

In: IEEE Transactions on Medical Imaging, Vol. 35, No. 4, 04.2016, p. 1127-1137.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Efficient small blob detection based on local convexity, intensity and shape information

AU - Zhang, Min

AU - Wu, Teresa

AU - Beeman, Scott C.

AU - Cullen-McEwen, Luise

AU - Bertram, John F.

AU - Charlton, Jennifer R.

AU - Baldelomar, Edwin

AU - Bennett, Kevin M.

PY - 2016/4

Y1 - 2016/4

N2 - The identification of small structures (blobs) from medical images to quantify clinically relevant features, such as size and shape, is important in many medical applications. One particular application explored here is the automated detection of kidney glomeruli after targeted contrast enhancement and magnetic resonance imaging. We propose a computationally efficient algorithm, termed the Hessian-based Difference of Gaussians (HDoG), to segment small blobs (e.g. glomeruli from kidney) from 3D medical images based on local convexity, intensity and shape information. The image is first smoothed and pre-segmented into small blob candidate regions based on local convexity. Two novel 3D regional features (regional blobness and regional flatness) are then extracted from the candidate regions. Together with regional intensity, the three features are used in an unsupervised learning algorithm for auto post-pruning. HDoG is first validated in a 2D form and compared with other three blob detectors from literature, which are generally for 2D images only. To test the detectability of blobs from 3D images, 240 sets of simulated images are rendered for scenarios mimicking the renal nephron distribution observed in contrast-enhanced, 3D MRI. The results show a satisfactory performance of HDoG in detecting large numbers of small blobs. Two sets of real kidney 3D MR images (6 rats, 3 human) are then used to validate the applicability of HDoG for glomeruli detection. By comparing MRI to stereological measurements, we verify that HDoG is a robust and efficient unsupervised technique for 3D blobs segmentation.

AB - The identification of small structures (blobs) from medical images to quantify clinically relevant features, such as size and shape, is important in many medical applications. One particular application explored here is the automated detection of kidney glomeruli after targeted contrast enhancement and magnetic resonance imaging. We propose a computationally efficient algorithm, termed the Hessian-based Difference of Gaussians (HDoG), to segment small blobs (e.g. glomeruli from kidney) from 3D medical images based on local convexity, intensity and shape information. The image is first smoothed and pre-segmented into small blob candidate regions based on local convexity. Two novel 3D regional features (regional blobness and regional flatness) are then extracted from the candidate regions. Together with regional intensity, the three features are used in an unsupervised learning algorithm for auto post-pruning. HDoG is first validated in a 2D form and compared with other three blob detectors from literature, which are generally for 2D images only. To test the detectability of blobs from 3D images, 240 sets of simulated images are rendered for scenarios mimicking the renal nephron distribution observed in contrast-enhanced, 3D MRI. The results show a satisfactory performance of HDoG in detecting large numbers of small blobs. Two sets of real kidney 3D MR images (6 rats, 3 human) are then used to validate the applicability of HDoG for glomeruli detection. By comparing MRI to stereological measurements, we verify that HDoG is a robust and efficient unsupervised technique for 3D blobs segmentation.

KW - kidney

KW - machine learning

KW - quantification and estimation

KW - segmentation

KW - shape analysis

UR - http://www.ncbi.nlm.nih.gov/pubmed/26685229

U2 - 10.1109/TMI.2015.2509463

DO - 10.1109/TMI.2015.2509463

M3 - Article

VL - 35

SP - 1127

EP - 1137

JO - IEEE Transactions on Medical Imaging

JF - IEEE Transactions on Medical Imaging

SN - 0278-0062

IS - 4

ER -