Cargando…
Deep feature loss to denoise OCT images using deep neural networks
Significance: Speckle noise is an inherent limitation of optical coherence tomography (OCT) images that makes clinical interpretation challenging. The recent emergence of deep learning could offer a reliable method to reduce noise in OCT images. Aim: We sought to investigate the use of deep features...
Autores principales: | , , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Society of Photo-Optical Instrumentation Engineers
2021
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8062795/ https://www.ncbi.nlm.nih.gov/pubmed/33893726 http://dx.doi.org/10.1117/1.JBO.26.4.046003 |
_version_ | 1783681836319244288 |
---|---|
author | Mehdizadeh, Maryam MacNish, Cara Xiao, Di Alonso-Caneiro, David Kugelman, Jason Bennamoun, Mohammed |
author_facet | Mehdizadeh, Maryam MacNish, Cara Xiao, Di Alonso-Caneiro, David Kugelman, Jason Bennamoun, Mohammed |
author_sort | Mehdizadeh, Maryam |
collection | PubMed |
description | Significance: Speckle noise is an inherent limitation of optical coherence tomography (OCT) images that makes clinical interpretation challenging. The recent emergence of deep learning could offer a reliable method to reduce noise in OCT images. Aim: We sought to investigate the use of deep features (VGG) to limit the effect of blurriness and increase perceptual sharpness and to evaluate its impact on the performance of OCT image denoising (DnCNN). Approach: Fifty-one macula-centered OCT pairs were used in training of the network. Another set of 20 OCT pair was used for testing. The DnCNN model was cascaded with a VGG network that acted as a perceptual loss function instead of the traditional losses of [Formula: see text] and [Formula: see text]. The VGG network remains fixed during the training process. We focused on the individual layers of the VGG-16 network to decipher the contribution of each distinctive layer as a loss function to produce denoised OCT images that were perceptually sharp and that preserved the faint features (retinal layer boundaries) essential for interpretation. The peak signal-to-noise ratio (PSNR), edge-preserving index, and no-reference image sharpness/blurriness [perceptual sharpness index (PSI), just noticeable blur (JNB), and spectral and spatial sharpness measure (S3)] metrics were used to compare deep feature losses with the traditional losses. Results: The deep feature loss produced images with high perceptual sharpness measures at the cost of less smoothness (PSNR) in OCT images. The deep feature loss outperformed the traditional losses ([Formula: see text] and [Formula: see text]) for all of the evaluation metrics except for PSNR. The PSI, S3, and JNB estimates of deep feature loss performance were 0.31, 0.30, and 16.53, respectively. For L1 and L2 losses performance, the PSI, [Formula: see text] , and JNB were 0.21 and 0.21, 0.17 and 0.16, and 14.46 and 14.34, respectively. Conclusions: We demonstrate the potential of deep feature loss in denoising OCT images. Our preliminary findings suggest research directions for further investigation. |
format | Online Article Text |
id | pubmed-8062795 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2021 |
publisher | Society of Photo-Optical Instrumentation Engineers |
record_format | MEDLINE/PubMed |
spelling | pubmed-80627952021-04-23 Deep feature loss to denoise OCT images using deep neural networks Mehdizadeh, Maryam MacNish, Cara Xiao, Di Alonso-Caneiro, David Kugelman, Jason Bennamoun, Mohammed J Biomed Opt Imaging Significance: Speckle noise is an inherent limitation of optical coherence tomography (OCT) images that makes clinical interpretation challenging. The recent emergence of deep learning could offer a reliable method to reduce noise in OCT images. Aim: We sought to investigate the use of deep features (VGG) to limit the effect of blurriness and increase perceptual sharpness and to evaluate its impact on the performance of OCT image denoising (DnCNN). Approach: Fifty-one macula-centered OCT pairs were used in training of the network. Another set of 20 OCT pair was used for testing. The DnCNN model was cascaded with a VGG network that acted as a perceptual loss function instead of the traditional losses of [Formula: see text] and [Formula: see text]. The VGG network remains fixed during the training process. We focused on the individual layers of the VGG-16 network to decipher the contribution of each distinctive layer as a loss function to produce denoised OCT images that were perceptually sharp and that preserved the faint features (retinal layer boundaries) essential for interpretation. The peak signal-to-noise ratio (PSNR), edge-preserving index, and no-reference image sharpness/blurriness [perceptual sharpness index (PSI), just noticeable blur (JNB), and spectral and spatial sharpness measure (S3)] metrics were used to compare deep feature losses with the traditional losses. Results: The deep feature loss produced images with high perceptual sharpness measures at the cost of less smoothness (PSNR) in OCT images. The deep feature loss outperformed the traditional losses ([Formula: see text] and [Formula: see text]) for all of the evaluation metrics except for PSNR. The PSI, S3, and JNB estimates of deep feature loss performance were 0.31, 0.30, and 16.53, respectively. For L1 and L2 losses performance, the PSI, [Formula: see text] , and JNB were 0.21 and 0.21, 0.17 and 0.16, and 14.46 and 14.34, respectively. Conclusions: We demonstrate the potential of deep feature loss in denoising OCT images. Our preliminary findings suggest research directions for further investigation. Society of Photo-Optical Instrumentation Engineers 2021-04-23 2021-04 /pmc/articles/PMC8062795/ /pubmed/33893726 http://dx.doi.org/10.1117/1.JBO.26.4.046003 Text en © 2021 The Authors https://creativecommons.org/licenses/by/4.0/Published by SPIE under a Creative Commons Attribution 4.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI. |
spellingShingle | Imaging Mehdizadeh, Maryam MacNish, Cara Xiao, Di Alonso-Caneiro, David Kugelman, Jason Bennamoun, Mohammed Deep feature loss to denoise OCT images using deep neural networks |
title | Deep feature loss to denoise OCT images using deep neural networks |
title_full | Deep feature loss to denoise OCT images using deep neural networks |
title_fullStr | Deep feature loss to denoise OCT images using deep neural networks |
title_full_unstemmed | Deep feature loss to denoise OCT images using deep neural networks |
title_short | Deep feature loss to denoise OCT images using deep neural networks |
title_sort | deep feature loss to denoise oct images using deep neural networks |
topic | Imaging |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8062795/ https://www.ncbi.nlm.nih.gov/pubmed/33893726 http://dx.doi.org/10.1117/1.JBO.26.4.046003 |
work_keys_str_mv | AT mehdizadehmaryam deepfeaturelosstodenoiseoctimagesusingdeepneuralnetworks AT macnishcara deepfeaturelosstodenoiseoctimagesusingdeepneuralnetworks AT xiaodi deepfeaturelosstodenoiseoctimagesusingdeepneuralnetworks AT alonsocaneirodavid deepfeaturelosstodenoiseoctimagesusingdeepneuralnetworks AT kugelmanjason deepfeaturelosstodenoiseoctimagesusingdeepneuralnetworks AT bennamounmohammed deepfeaturelosstodenoiseoctimagesusingdeepneuralnetworks |