A Region Segmentation Method on 2-D Vessel Optical Coherence Tomography Images

Abstract

This paper describes a novel region segmentation method designed to avoid complications of the threshold process used in traditional segmentation methods in 2-D optical coherence tomography (OCT) images. Analysis of the layers and regions in OCT images is used to diagnose the presence of cancer and identify the stage of the cancer if present. However, scattering during OCT images generates a speckle effect and creates diffusion problems which are also captured; these problems cause traditional image processing methods such as the Canny edge and Otsu methods to fail in finding the proper layer and region edges. The proposed method uses the mean value and an enhanced-fuzzy-c-mean algorithm to cluster pixels in 2-D OCT images and find the edge between different clustered regions. Low-resolution vessel OCT and high-resolution oral cancer OCT images are tested in the experiment, and the experimental results show that the proposed method performs with more robust and accurate segmentation results than does the overcomplete-wavelet-frame-based fractal signature method.