Segmentation of Corneal Optical Coherence Tomography Images Using Randomized Hough Transform

摘要

Measuring the thickness of different corneal microlayers is important for the diagnosis of common corneal eyediseases such as dry eye, keratoconus, Fuchs endothelial dystrophy, and corneal graft rejection. High resolutioncorneal images, obtained using optical coherence tomography (OCT), made it possible to measure the thickness ofdifferent corneal microlayers in vivo. The manual segmentation of these images is subjective and time consuming.Therefore, automatic segmentation is necessary. Several methods were proposed for segmenting corneal OCTimages, but none of these methods segment all the microlayer interfaces and they are not robust. In addition,the lack of a large annotated database of corneal OCT images impedes the application of machine learningmethods such as deep learning which proves to be very powerful. In this paper, we present a new cornealOCT image segmentation algorithm using Randomized Hough Transform. To the best of our knowledge, wedeveloped the first automatic segmentation method for the six corneal microlayer interfaces. The proposedmethod includes a robust estimate of relative distances of inner corneal interfaces with respect to outer cornealinterfaces. Also, it handles properly the correct ordering and the non-intersection of corneal microlayer interfaces.The proposed method was tested on 15 corneal OCT images that were randomly selected. OCT images weremanually segmented by two trained operators for comparison. Comparison with the manual segmentation showsthat the proposed method has mean segmentation error of 3:77-4:25 pixels across all interfaces which correspondsto 5:66 - 6:38μm. The mean segmentation error between the two manual operators is 4:07 - 4:71 pixels, whichcorresponds to 6:11 - 7:07μm. The proposed method takes a mean time of 2:59 - 0:06 seconds to segment sixcorneal interfaces.