|
|||||
|
|
Reduction of speckle noise from optical coherence tomography images using multi-frame weighted nuclear norm minimization method |
|
| Authors: | Damber Thapa Kaamran Raahemifar Vasudevan Lakshminarayanan |
| Affiliation: | 1. School of Optometry and Vision Science, University of Waterloo, Waterloo, Canadadthapa@uwaterloo.ca;3. Department of Electrical and Computer Engineering, Ryerson University, Toronto, Canada;4. School of Optometry and Vision Science, University of Waterloo, Waterloo, Canada;5. Department of Physics and Electrical and Computer Engineering, University of Waterloo, Waterloo, Canada;6. Department of Physics, University of Michigan, Ann Arbor, MI, USA |
| Abstract: | In this paper, we propose a speckle noise reduction method for spectral-domain optical coherence tomography (SD-OCT) images called multi-frame weighted nuclear norm minimization (MWNNM). This method is a direct extension of weighted nuclear norm minimization (WNNM) in the multi-frame framework since an adequately denoised image could not be achieved with single-frame denoising methods. The MWNNM method exploits multiple B-scans collected from a small area of a SD-OCT volumetric image, and then denoises and averages them together to obtain a high signal-to-noise ratio B-scan. The results show that the image quality metrics obtained by denoising and averaging only five nearby B-scans with MWNNM method is considerably better than those of the average image obtained by registering and averaging 40 azimuthally repeated B-scans. |
| Keywords: | denoising reconstruction nuclear norm minimization singular value decomposition image processing ophthalmology retinal images optical coherence tomography |