Optical Coherence Tomography Denoising by Means of a Fourier Butterworth Filter-Based Approach

Abstract

[Abstract]: Optical Coherence Tomography (OCT) is affected by ubiquitous speckle noise that difficult the visualization and analysis of the retinal structures. Any denoising strategy should be able to remove efficiently the noise as well as preserves clinical information contained in the images. This information is crucial to analyses the retinal layer tissue that allows the posterior analysis and recognition of relevant diseases as macular edema or diabetic retinopathy. To address this issue, a method based on the Fourier Butterworth filter combined with a contrast enhancement and a histogram regularization was developed in order to reduce the speckle noise in OCT retinal images. The proposed method was validated using 45 OCT retinal images organized into 3 groups of noise degree, comparing the results with the performance of representative methods of the state-of-the-art. The validation and comparison were made through three quantitative metrics: Signal-to-Noise Ratio (SNR), Contrast-to-Noise Ratio (CNR) and average Effective Number of Looks (ENL). The experimental results showed that the proposed method offered satisfactory results, outperforming the results of the other methods by the achievement of a SNR of 7.04 dB and a CNR of 14.08 dB better than the second best filter, respectively, for the whole group of OCT retinal images.