A Variational Model Based on 3D Transformation to Remove the Speckle Noise in OCT Images

This paper presents a new variational model through designing a novel regularizing term based on Block Matching and 3-D Filtering (BM3D), which is efficient to denoise speckled images in optical coherence tomography (OCT). In the proposed algorithm, speckle noise is modeled as Rayleigh distribution, which is usually made additive by means of a logarithmic transformation in other digital filtering methods. At last, the model is applied to the denoising of a Port wine stain (PWS) skin OCT image and the experimental results demonstrate that the proposed algorithm is better than some other filtering methods, it is more consistent in the characteristic of OCT speckle noise and is superior in preserving small scattering structures.