Self-supervised Blind2Unblind deep learning scheme for OCT speckle reductions
As a low-coherence interferometry-based imaging modality, optical coherence tomography (OCT) inevitably suffers from the influence of speckles originating from multiply scattered photons. Speckles hide tissue microstructures and degrade the accuracy of disease diagnoses, which thus hinder OCT clinic...
Saved in:
Main Authors: | , , , , , , , |
---|---|
Other Authors: | |
Format: | Article |
Language: | English |
Published: |
2023
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/171481 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
id |
sg-ntu-dr.10356-171481 |
---|---|
record_format |
dspace |
spelling |
sg-ntu-dr.10356-1714812023-10-27T15:40:09Z Self-supervised Blind2Unblind deep learning scheme for OCT speckle reductions Yu, Xiaojun Ge, Chenkun Li, Mingshuai Yuan, Miao Liu, Linbo Mo, Jianhua Shum, Perry Ping Chen, Jinna School of Electrical and Electronic Engineering Engineering::Electrical and electronic engineering Optical Coherence Tomography Images As a low-coherence interferometry-based imaging modality, optical coherence tomography (OCT) inevitably suffers from the influence of speckles originating from multiply scattered photons. Speckles hide tissue microstructures and degrade the accuracy of disease diagnoses, which thus hinder OCT clinical applications. Various methods have been proposed to address such an issue, yet they suffer either from the heavy computational load, or the lack of high-quality clean images prior, or both. In this paper, a novel self-supervised deep learning scheme, namely, Blind2Unblind network with refinement strategy (B2Unet), is proposed for OCT speckle reduction with a single noisy image only. Specifically, the overall B2Unet network architecture is presented first, and then, a global-aware mask mapper together with a loss function are devised to improve image perception and optimize sampled mask mapper blind spots, respectively. To make the blind spots visible to B2Unet, a new re-visible loss is also designed, and its convergence is discussed with the speckle properties being considered. Extensive experiments with different OCT image datasets are finally conducted to compare B2Unet with those state-of-the-art existing methods. Both qualitative and quantitative results convincingly demonstrate that B2Unet outperforms the state-of-the-art model-based and fully supervised deep-learning methods, and it is robust and capable of effectively suppressing speckles while preserving the important tissue micro-structures in OCT images in different cases. Published version National Natural Science Foundation of China (62220106006); Basic and Applied Basic Research Foundation of Guangdong Province (2021B1515120013); Key Research and Development Projects of Shaanxi Province (2021SF-342); Key Research Project of Shaanxi Higher Education Teaching Reform (21BG005). 2023-10-26T03:24:07Z 2023-10-26T03:24:07Z 2023 Journal Article Yu, X., Ge, C., Li, M., Yuan, M., Liu, L., Mo, J., Shum, P. P. & Chen, J. (2023). Self-supervised Blind2Unblind deep learning scheme for OCT speckle reductions. Biomedical Optics Express, 14(6), 2773-2795. https://dx.doi.org/10.1364/BOE.481870 2156-7085 https://hdl.handle.net/10356/171481 10.1364/BOE.481870 37342690 2-s2.0-85162019281 6 14 2773 2795 en Biomedical Optics Express © 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement. application/pdf |
institution |
Nanyang Technological University |
building |
NTU Library |
continent |
Asia |
country |
Singapore Singapore |
content_provider |
NTU Library |
collection |
DR-NTU |
language |
English |
topic |
Engineering::Electrical and electronic engineering Optical Coherence Tomography Images |
spellingShingle |
Engineering::Electrical and electronic engineering Optical Coherence Tomography Images Yu, Xiaojun Ge, Chenkun Li, Mingshuai Yuan, Miao Liu, Linbo Mo, Jianhua Shum, Perry Ping Chen, Jinna Self-supervised Blind2Unblind deep learning scheme for OCT speckle reductions |
description |
As a low-coherence interferometry-based imaging modality, optical coherence tomography (OCT) inevitably suffers from the influence of speckles originating from multiply scattered photons. Speckles hide tissue microstructures and degrade the accuracy of disease diagnoses, which thus hinder OCT clinical applications. Various methods have been proposed to address such an issue, yet they suffer either from the heavy computational load, or the lack of high-quality clean images prior, or both. In this paper, a novel self-supervised deep learning scheme, namely, Blind2Unblind network with refinement strategy (B2Unet), is proposed for OCT speckle reduction with a single noisy image only. Specifically, the overall B2Unet network architecture is presented first, and then, a global-aware mask mapper together with a loss function are devised to improve image perception and optimize sampled mask mapper blind spots, respectively. To make the blind spots visible to B2Unet, a new re-visible loss is also designed, and its convergence is discussed with the speckle properties being considered. Extensive experiments with different OCT image datasets are finally conducted to compare B2Unet with those state-of-the-art existing methods. Both qualitative and quantitative results convincingly demonstrate that B2Unet outperforms the state-of-the-art model-based and fully supervised deep-learning methods, and it is robust and capable of effectively suppressing speckles while preserving the important tissue micro-structures in OCT images in different cases. |
author2 |
School of Electrical and Electronic Engineering |
author_facet |
School of Electrical and Electronic Engineering Yu, Xiaojun Ge, Chenkun Li, Mingshuai Yuan, Miao Liu, Linbo Mo, Jianhua Shum, Perry Ping Chen, Jinna |
format |
Article |
author |
Yu, Xiaojun Ge, Chenkun Li, Mingshuai Yuan, Miao Liu, Linbo Mo, Jianhua Shum, Perry Ping Chen, Jinna |
author_sort |
Yu, Xiaojun |
title |
Self-supervised Blind2Unblind deep learning scheme for OCT speckle reductions |
title_short |
Self-supervised Blind2Unblind deep learning scheme for OCT speckle reductions |
title_full |
Self-supervised Blind2Unblind deep learning scheme for OCT speckle reductions |
title_fullStr |
Self-supervised Blind2Unblind deep learning scheme for OCT speckle reductions |
title_full_unstemmed |
Self-supervised Blind2Unblind deep learning scheme for OCT speckle reductions |
title_sort |
self-supervised blind2unblind deep learning scheme for oct speckle reductions |
publishDate |
2023 |
url |
https://hdl.handle.net/10356/171481 |
_version_ |
1781793715069648896 |