Layer Structures Localization in Optical Coherence Tomography Images

The existence of noise and speckle in optical coherence tomography (OCT) images challenges the extraction of layer structures. As an effort to answer the challenges, an improved coherence enhancing diffusion (CED) algorithm for preprocessing OCT images is presented in this paper. The improved algorithm could keep the locality of layer structures in OCT images better, therefore, locality of layer structures could be detected with good accuracy in preprocessed OCT images. The locality of some specific layer structures found in preprocessed images could be used as the benchmark to optimize OCT images, the application in OCT images of retina is such an example. Experiment results show that the above optimization is clinical helpful.