A Joint Model for Macular Edema Analysis in Optical Coherence Tomography Images Based on Image Enhancement and Segmentation

Algorithm 1

A joint model for macular edema segmentation in OCT images.

Input: OCT images

Output: segmentation results of macular edema in OCT images

Step 1: the quality of input OCT image was improved using the bioinspired algorithm as shown below:

(i) Denoise the input OCT images by GF with kernel

(ii) Preserve the edge of denoised OCT image by structure-preserving guided retinal image filtering (SGRIF)

(iii) Enhance the contrast of processed OCT image by single-scale Retinex (SSR)

Step 2: taking the enhanced OCT image as the input data, the edema region was segmented by SBGFRLS algorithm as shown below:

(i) Initialize the level set function

(ii) Compute using equation ((11)), which is the inside and outside curve of edema region

(iii) Compute the signed pressure function using by equation (10)

(iv) Evolve the level set function according to equation (9)

(v) Regularize the level set function with a Gaussian filter, where if , let ; otherwise,

(vi) Segment the edema region by iterative computations from (ii) to (v).