<scp>N2NSR‐OCT</scp>: Simultaneous denoising and super‐resolution in optical coherence tomography images using semisupervised deep learning

Qiu, Bin; You, Yunfei; Huang, Zhiyu; Meng, Xiangxi; Jiang, Zhe; Zhou, Chuanqing; Liu, Gangjun; Yang, Kun; Ren, Qiushi; Lu, Yanye

doi:10.1002/jbio.202000282

Cited by 34 publications

(21 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…That is, RE-OCT can only ‘boost’ the signal that is originally present (i.e., non-zero SNR), but not ‘create’ new signal that did not exist. Retrieving the spatial frequencies beyond the physical aperture of the optical system is beyond the scope of RE-OCT; this would require analytic continuation 64 and could potentially be achieved by some super-resolution and/or deep learning approaches 65 , 66 .…”

Section: Resultsmentioning

confidence: 99%

Resolution-enhanced OCT and expanded framework of information capacity and resolution in coherent imaging

Leartprapun

Adie

2021

Sci Rep

View full text Add to dashboard Cite

Spatial resolution in conventional optical microscopy has traditionally been treated as a fixed parameter of the optical system. Here, we present an approach to enhance transverse resolution in beam-scanned optical coherence tomography (OCT) beyond its aberration-free resolution limit, without any modification to the optical system. Based on the theorem of invariance of information capacity, resolution-enhanced (RE)-OCT navigates the exchange of information between resolution and signal-to-noise ratio (SNR) by exploiting efficient noise suppression via coherent averaging and a simple computational bandwidth expansion procedure. We demonstrate a resolution enhancement of 1.5 × relative to the aberration-free limit while maintaining comparable SNR in silicone phantom. We show that RE-OCT can significantly enhance the visualization of fine microstructural features in collagen gel and ex vivo mouse brain. Beyond RE-OCT, our analysis in the spatial-frequency domain leads to an expanded framework of information capacity and resolution in coherent imaging that contributes new implications to the theory of coherent imaging. RE-OCT can be readily implemented on most OCT systems worldwide, immediately unlocking information that is beyond their current imaging capabilities, and so has the potential for widespread impact in the numerous areas in which OCT is utilized, including the basic sciences and translational medicine.

show abstract

Section: Resultsmentioning

confidence: 99%

Resolution-enhanced OCT and expanded framework of information capacity and resolution in coherent imaging

Leartprapun

Adie

2021

Sci Rep

View full text Add to dashboard Cite

show abstract

“…2) DL-based methods for OCT image reconstruction: Variations of U-Net have been adopted in reconstructing OCT images [12], [13], [16], [20]- [22]. Hao et al reconstructed high quality OCT images from their low bit-depth counterparts using a U-Net based architecture [21].…”

Section: B Oct Image Reconstruction 1) None-dl Methods For Oct Image ...mentioning

confidence: 99%

“…Also with the U-Net architecture, Liang et al enhanced the resolution while recovering realistic speckles in the OCT images [13]. Further, Qiu et al achieved simultaneously denoising and reconstruction of OCT images using a U-Net based network [22]. The U-Net based networks have been used in reconstructing OCT images compressed in spectral domain.…”

Section: B Oct Image Reconstruction 1) None-dl Methods For Oct Image ...mentioning

confidence: 99%

Multi-scale reconstruction of undersampled spectral-spatial OCT data for coronary imaging using deep learning

Li¹,

Cao²,

Liu³

et al. 2022

Preprint

View full text Add to dashboard Cite

Coronary artery disease (CAD) is a cardiovascular condition with high morbidity and mortality. Intravascular optical coherence tomography (IVOCT) has been considered as an optimal imagining system for the diagnosis and treatment of CAD. Constrained by Nyquist theorem, dense sampling in IVOCT attains high resolving power to delineate cellular structures/features. There is a trade-off between high spatial resolution and fast scanning rate for coronary imaging. In this paper, we propose a viable spectral-spatial acquisition method that down-scales the sampling process in both spectral and spatial domain while maintaining high quality in image reconstruction. The down-scaling schedule boosts data acquisition speed without any hardware modifications. Additionally, we propose a unified multi-scale reconstruction framework, namely Multiscale-Spectral-Spatial-Magnification Network (MSSMN), to resolve highly down-scaled (compressed) OCT images with flexible magnification factors. We incorporate the proposed methods into Spectral Domain OCT (SD-OCT) imaging of human coronary samples with clinical features such as stent and calcified lesions. Our experimental results demonstrate that spectral-spatial downscaled data can be better reconstructed than data that is downscaled solely in either spectral or spatial domain. Moreover, we observe better reconstruction performance using MSSMN than using existing reconstruction methods. Our acquisition method and multi-scale reconstruction framework, in combination, may allow faster SD-OCT inspection with high resolution during coronary intervention.

show abstract

“…In medical image field, encoder-decoder networkbased models [28][29][30][31] and GAN-based models [32][33][34] are two important implements for DNN-based applications. For encoder-decoder network-based models, the hierarchical features can be extracted by using the stack of convolutional layers so that the encoder-decoder network-based models were widely used in medical image processing and analysis tasks, including image denoising [31], super-resolution [28,35], classification [36] and segmentation [30]. According to the characteristic of information stream in the models, the encoderdecoder network-based models could be further divided into U-shaped model, multi-information stream model and straight-information stream model.…”

Section: Structure Of Dnnsmentioning

confidence: 99%

“…One of the most important superiorities of UNet is that the skip-connection part of different level is employed, which assembles the feature representations across multiple semantic scales and prevents information loss. Then due to the outstanding representation capability of network, UNet was further extended to numerous end-to-end medical image translation tasks [35] and was further modified to U-shaped models, that is, UNet++ and attention U-Net. Hence, a modified UNet network combined with N2N strategy…”

Section: U-shaped Modelmentioning

confidence: 99%

Comparative study of deep neural networks with unsupervised Noise2Noise strategy for noise reduction of optical coherence tomography images

Qiu

Zeng

Meng

et al. 2021

Journal of Biophotonics

Self Cite

View full text Add to dashboard Cite

As a powerful diagnostic tool, optical coherence tomography (OCT) has been widely used in various clinical setting. However, OCT images are susceptible to inherent speckle noise that may contaminate subtle structure information, due to low-coherence interferometric imaging procedure. Many supervised learningbased models have achieved impressive performance in reducing speckle noise of OCT images trained with a large number of noisy-clean paired OCT images, which are not commonly feasible in clinical practice. In this article, we conducted a comparative study to investigate the denoising performance of OCT images over different deep neural networks through an unsupervised Noise2Noise (N2N) strategy, which only trained with noisy OCT samples. Four representative network architectures including U-shaped model, multi-information stream model, straightinformation stream model and GAN-based model were investigated on an OCT image dataset acquired from healthy human eyes. The results demonstrated all four unsupervised N2N models offered denoised OCT images with aperformance comparable with that of supervised learning models, illustrating the effectiveness of unsupervised N2N models in denoising OCT images. Furthermore, U-shaped models and GAN-based models using UNet network as generator are two preferred and suitable architectures for reducing speckle noise of OCT images and preserving fine structure information of retinal layers under unsupervised N2N circumstances.

show abstract

N2NSR‐OCT: Simultaneous denoising and super‐resolution in optical coherence tomography images using semisupervised deep learning

Cited by 34 publications

References 28 publications

Resolution-enhanced OCT and expanded framework of information capacity and resolution in coherent imaging

Resolution-enhanced OCT and expanded framework of information capacity and resolution in coherent imaging

Multi-scale reconstruction of undersampled spectral-spatial OCT data for coronary imaging using deep learning

Comparative study of deep neural networks with unsupervised Noise2Noise strategy for noise reduction of optical coherence tomography images

Contact Info

Product

Resources

About