TY - GEN
T1 - Speckle noise reduction in OCT and projection images using hybrid wavelet thresholding
AU - Sui, X.
AU - Ishikawa, H.
AU - Selesnick, I. W.
AU - Wollstein, G.
AU - Schuman, J. S.
N1 - Funding Information:
ACKNOWLEDGMENT This work is partially supported by the National Science Foundation under Grant No. CCF-1525398.
Funding Information:
This work is partially supported by the National Science Foundation under Grant No. CCF-1525398.
Publisher Copyright:
© 2018 IEEE.
PY - 2018/7/2
Y1 - 2018/7/2
N2 - Speckle noise in optical coherence tomography (OCT) images is a granular noise that inherently exists and degrades the image quality. The challenge of conventional denoising methods is to distinguish the informational pattern from the speckle noise. In this paper we present a novel method for speckle noise reduction in OCT volumes, where the corresponding en face representation, which produces frontal sections of retinal layers and is relatively free of speckle, is considered as a reference. The proposed method estimates the anatomical structures by solving a constrained optimization problem that combines wavelet-domain sparsity and total variation (wavelet-TV) regularization to preserve the edges of retinal layers and to alleviate artifacts introduced by pure wavelet thresholding. Denoising performance is evaluated through the signal to noise ratio (SNR) and the contrast to noise ratio (CNR). The volumes processed by the proposed method show notable reduction of speckle without losing details in both en face and cross-sectional images.
AB - Speckle noise in optical coherence tomography (OCT) images is a granular noise that inherently exists and degrades the image quality. The challenge of conventional denoising methods is to distinguish the informational pattern from the speckle noise. In this paper we present a novel method for speckle noise reduction in OCT volumes, where the corresponding en face representation, which produces frontal sections of retinal layers and is relatively free of speckle, is considered as a reference. The proposed method estimates the anatomical structures by solving a constrained optimization problem that combines wavelet-domain sparsity and total variation (wavelet-TV) regularization to preserve the edges of retinal layers and to alleviate artifacts introduced by pure wavelet thresholding. Denoising performance is evaluated through the signal to noise ratio (SNR) and the contrast to noise ratio (CNR). The volumes processed by the proposed method show notable reduction of speckle without losing details in both en face and cross-sectional images.
UR - http://www.scopus.com/inward/record.url?scp=85062098025&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85062098025&partnerID=8YFLogxK
U2 - 10.1109/SPMB.2018.8615623
DO - 10.1109/SPMB.2018.8615623
M3 - Conference contribution
AN - SCOPUS:85062098025
T3 - 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings
BT - 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2018 IEEE Signal Processing in Medicine and Biology Symposium, SPMB 2018
Y2 - 1 December 2018
ER -