An intelligent speckle reduction algorithm for optical coherence tomography images

Abstract

Optical Coherence Tomography (OCT) offers three dimensional images of tissue microstructures. Although OCT imaging offers a promising high resolution method, due to the low coherent light source used in the configuration of OCT, OCT images suffers from an artefact called, speckle. Speckle deteriorates the image quality and effects image analysis algorithm such as segmentation and pattern recognition. We present a novel and intelligent speckle reduction algorithm to reduce speckle based on an ensemble framework of Multi-Layer Perceptron (MLP) neural networks. We tested the algorithm on images of retina obtained from a spectrometer-based Fourier-domain OCT system operating at 890 nm, and observed considerable improvement in the signal-to-noise ratio and contrast of the images.