TY - JOUR
T1 - Retinal optical coherence tomography image enhancement via deep learning
AU - Halupka, Kerry J.
AU - Antony, Bhavna J.
AU - Lee, Matthew H.
AU - Lucy, Katie A.
AU - Rai, Ravneet S.
AU - Ishikawa, Hiroshi
AU - Wollstein, Gadi
AU - Schuman, Joel S.
AU - Garnavi, Rahil
N1 - Publisher Copyright:
© 2018 Optical Society of America.
PY - 2018/12/1
Y1 - 2018/12/1
N2 - Optical coherence tomography (OCT) images of the retina are a powerful tool for diagnosing and monitoring eye disease. However, they are plagued by speckle noise, which reduces image quality and reliability of assessment. This paper introduces a novel speckle reduction method inspired by the recent successes of deep learning in medical imaging. We present two versions of the network to reflect the needs and preferences of different end-users. Specifically, we train a convolution neural network to denoise cross-sections from OCT volumes of healthy eyes using either (1) mean-squared error, or (2) a generative adversarial network (GAN) with Wasserstein distance and perceptual similarity. We then interrogate the success of both methods with extensive quantitative and qualitative metrics on cross-sections from both healthy and glaucomatous eyes. The results show that the former approach provides state-of-the-art improvement in quantitative metrics such as PSNR and SSIM, and aids layer segmentation. However, the latter approach, which puts more weight on visual perception, outperformed for qualitative comparisons based on accuracy, clarity, and personal preference. Overall, our results demonstrate the effectiveness and efficiency of a deep learning approach to denoising OCT images, while maintaining subtle details in the images.
AB - Optical coherence tomography (OCT) images of the retina are a powerful tool for diagnosing and monitoring eye disease. However, they are plagued by speckle noise, which reduces image quality and reliability of assessment. This paper introduces a novel speckle reduction method inspired by the recent successes of deep learning in medical imaging. We present two versions of the network to reflect the needs and preferences of different end-users. Specifically, we train a convolution neural network to denoise cross-sections from OCT volumes of healthy eyes using either (1) mean-squared error, or (2) a generative adversarial network (GAN) with Wasserstein distance and perceptual similarity. We then interrogate the success of both methods with extensive quantitative and qualitative metrics on cross-sections from both healthy and glaucomatous eyes. The results show that the former approach provides state-of-the-art improvement in quantitative metrics such as PSNR and SSIM, and aids layer segmentation. However, the latter approach, which puts more weight on visual perception, outperformed for qualitative comparisons based on accuracy, clarity, and personal preference. Overall, our results demonstrate the effectiveness and efficiency of a deep learning approach to denoising OCT images, while maintaining subtle details in the images.
UR - http://www.scopus.com/inward/record.url?scp=85057793803&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85057793803&partnerID=8YFLogxK
U2 - 10.1364/BOE.9.006205
DO - 10.1364/BOE.9.006205
M3 - Article
AN - SCOPUS:85057793803
SN - 2156-7085
VL - 9
SP - 6205
EP - 6221
JO - Biomedical Optics Express
JF - Biomedical Optics Express
IS - 12
M1 - #342299
ER -