Cluster-based filtering framework for speckle reduction in OCT images

• DOI: 10.1364/boe.9.006359
• Publication Date: 2018-11-19T20:48:28Z
• Authors: M. Hossein Eybposh; Zahra Turani; Darius Mehregan; Mohammadreza Nasi...
• Keywords: Speckle noise; Smoothing

Optical coherence tomography (OCT) has become a popular modality in the dermatology discipline due to its moderate resolution and penetration depth. OCT images, however, contain grainy pattern called speckle. To date, variety of filtering techniques have been introduced reduce speckle images. However, further improvement is required edge smoothing deterioration small structures images after despeckling. In this manuscript, we present novel cluster-based reduction framework (CSRF) that consists clustering method, followed by despeckling method. Since edges are borders two adjacent clusters, proposed leaves intact. Moreover, multiplicative noise could be modeled as additive each cluster. evaluate performance CSRF demonstrate generic nature, namely k-means (KM), and, pixelwise algorithms, including Lee filter (LF) adaptive Wiener (AWF), used. The results indicate significantly improves algorithms. These improvements evaluated on healthy human skin vivo using numerical assessment measures signal-to-noise ratio (SNR), structural similarity index (SSIM).