Three dimensional data-driven multi scale atomic representation of optical coherence tomography

Raheleh Kafieh, Hossein Rabbani, Ivan Selesnick

Research output: Contribution to journalArticle

Abstract

In this paper, we discuss about applications of different methods for decomposing a signal over elementary waveforms chosen in a family called a dictionary (atomic representations) in optical coherence tomography (OCT). If the representation is learned from the data, a nonparametric dictionary is defined with three fundamental properties of being data-driven, applicability on 3D, and working in multi-scale, which make it appropriate for processing of OCT images. We discuss about application of such representations including complex wavelet based K-SVD, and diffusion wavelets on OCT data. We introduce complex wavelet based K-SVD to take advantage of adaptability in dictionary learning methods to improve the performance of simple dual tree complex wavelets in speckle reduction of OCT datasets in 2D and 3D. The algorithm is evaluated on 144 randomly selected slices from twelve 3D OCTs taken by Topcon 3D OCT-1000 and Cirrus Zeiss Meditec. Improvement of contrast to noise ratio (CNR) (from 0.9 to 11.91 and from 3.09 to 88.9, respectively) is achieved. Furthermore, two approaches are proposed for image segmentation using diffusion. The first method is designing a competition between extended basis functions at each level and the second approach is defining a new distance for each level and clustering based on such distances. A combined algorithm, based on these two methods is then proposed for segmentation of retinal OCTs, which is able to localize 12 boundaries with unsigned border positioning error of 9.22 ± 3.05 μm, on a test set of 20 slices selected from 13 3D OCTs.

Original languageEnglish (US)
Article number6967823
Pages (from-to)1042-1062
Number of pages21
JournalIEEE Transactions on Medical Imaging
Volume34
Issue number5
DOIs
StatePublished - May 1 2015

Fingerprint

Optical tomography
Optical Coherence Tomography
Glossaries
Singular value decomposition
Speckle
Image segmentation
Cluster Analysis
Noise
Learning
Processing

Keywords

  • Denoising
  • dictionary learning
  • diffusion wavelet
  • optical coherence tomography
  • segmentation
  • three-dimensional (3D) complex wavelet transform

ASJC Scopus subject areas

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

Cite this

Three dimensional data-driven multi scale atomic representation of optical coherence tomography. / Kafieh, Raheleh; Rabbani, Hossein; Selesnick, Ivan.

In: IEEE Transactions on Medical Imaging, Vol. 34, No. 5, 6967823, 01.05.2015, p. 1042-1062.

Research output: Contribution to journalArticle

@article{3f87cdfb3eaf43f0acd9f54bfdc42aad,
title = "Three dimensional data-driven multi scale atomic representation of optical coherence tomography",
abstract = "In this paper, we discuss about applications of different methods for decomposing a signal over elementary waveforms chosen in a family called a dictionary (atomic representations) in optical coherence tomography (OCT). If the representation is learned from the data, a nonparametric dictionary is defined with three fundamental properties of being data-driven, applicability on 3D, and working in multi-scale, which make it appropriate for processing of OCT images. We discuss about application of such representations including complex wavelet based K-SVD, and diffusion wavelets on OCT data. We introduce complex wavelet based K-SVD to take advantage of adaptability in dictionary learning methods to improve the performance of simple dual tree complex wavelets in speckle reduction of OCT datasets in 2D and 3D. The algorithm is evaluated on 144 randomly selected slices from twelve 3D OCTs taken by Topcon 3D OCT-1000 and Cirrus Zeiss Meditec. Improvement of contrast to noise ratio (CNR) (from 0.9 to 11.91 and from 3.09 to 88.9, respectively) is achieved. Furthermore, two approaches are proposed for image segmentation using diffusion. The first method is designing a competition between extended basis functions at each level and the second approach is defining a new distance for each level and clustering based on such distances. A combined algorithm, based on these two methods is then proposed for segmentation of retinal OCTs, which is able to localize 12 boundaries with unsigned border positioning error of 9.22 ± 3.05 μm, on a test set of 20 slices selected from 13 3D OCTs.",
keywords = "Denoising, dictionary learning, diffusion wavelet, optical coherence tomography, segmentation, three-dimensional (3D) complex wavelet transform",
author = "Raheleh Kafieh and Hossein Rabbani and Ivan Selesnick",
year = "2015",
month = "5",
day = "1",
doi = "10.1109/TMI.2014.2374354",
language = "English (US)",
volume = "34",
pages = "1042--1062",
journal = "IEEE Transactions on Medical Imaging",
issn = "0278-0062",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

TY - JOUR

T1 - Three dimensional data-driven multi scale atomic representation of optical coherence tomography

AU - Kafieh, Raheleh

AU - Rabbani, Hossein

AU - Selesnick, Ivan

PY - 2015/5/1

Y1 - 2015/5/1

N2 - In this paper, we discuss about applications of different methods for decomposing a signal over elementary waveforms chosen in a family called a dictionary (atomic representations) in optical coherence tomography (OCT). If the representation is learned from the data, a nonparametric dictionary is defined with three fundamental properties of being data-driven, applicability on 3D, and working in multi-scale, which make it appropriate for processing of OCT images. We discuss about application of such representations including complex wavelet based K-SVD, and diffusion wavelets on OCT data. We introduce complex wavelet based K-SVD to take advantage of adaptability in dictionary learning methods to improve the performance of simple dual tree complex wavelets in speckle reduction of OCT datasets in 2D and 3D. The algorithm is evaluated on 144 randomly selected slices from twelve 3D OCTs taken by Topcon 3D OCT-1000 and Cirrus Zeiss Meditec. Improvement of contrast to noise ratio (CNR) (from 0.9 to 11.91 and from 3.09 to 88.9, respectively) is achieved. Furthermore, two approaches are proposed for image segmentation using diffusion. The first method is designing a competition between extended basis functions at each level and the second approach is defining a new distance for each level and clustering based on such distances. A combined algorithm, based on these two methods is then proposed for segmentation of retinal OCTs, which is able to localize 12 boundaries with unsigned border positioning error of 9.22 ± 3.05 μm, on a test set of 20 slices selected from 13 3D OCTs.

AB - In this paper, we discuss about applications of different methods for decomposing a signal over elementary waveforms chosen in a family called a dictionary (atomic representations) in optical coherence tomography (OCT). If the representation is learned from the data, a nonparametric dictionary is defined with three fundamental properties of being data-driven, applicability on 3D, and working in multi-scale, which make it appropriate for processing of OCT images. We discuss about application of such representations including complex wavelet based K-SVD, and diffusion wavelets on OCT data. We introduce complex wavelet based K-SVD to take advantage of adaptability in dictionary learning methods to improve the performance of simple dual tree complex wavelets in speckle reduction of OCT datasets in 2D and 3D. The algorithm is evaluated on 144 randomly selected slices from twelve 3D OCTs taken by Topcon 3D OCT-1000 and Cirrus Zeiss Meditec. Improvement of contrast to noise ratio (CNR) (from 0.9 to 11.91 and from 3.09 to 88.9, respectively) is achieved. Furthermore, two approaches are proposed for image segmentation using diffusion. The first method is designing a competition between extended basis functions at each level and the second approach is defining a new distance for each level and clustering based on such distances. A combined algorithm, based on these two methods is then proposed for segmentation of retinal OCTs, which is able to localize 12 boundaries with unsigned border positioning error of 9.22 ± 3.05 μm, on a test set of 20 slices selected from 13 3D OCTs.

KW - Denoising

KW - dictionary learning

KW - diffusion wavelet

KW - optical coherence tomography

KW - segmentation

KW - three-dimensional (3D) complex wavelet transform

UR - http://www.scopus.com/inward/record.url?scp=84929458566&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84929458566&partnerID=8YFLogxK

U2 - 10.1109/TMI.2014.2374354

DO - 10.1109/TMI.2014.2374354

M3 - Article

VL - 34

SP - 1042

EP - 1062

JO - IEEE Transactions on Medical Imaging

JF - IEEE Transactions on Medical Imaging

SN - 0278-0062

IS - 5

M1 - 6967823

ER -