Fluid Segmentation in Retinal OCT images with Pigment Epithelial Detachment

For the assessment of Pigment Epithelial Detachment(PED), optical coherence tomography (OCT) must automatically and precisely segment the fluid related subretinal fluid . An algorithm for guided fluid segmentation in two stages is suggested. To get the primary curve of the fluid part for the Graph cut approach, the candidate fluid region is automatic determined in the first step. The Gaussian pre-fitting approach is suggested to segment sub-retinal-fluid in the second stage. The suggested algorithm is evaluated using the testing data set, which consists of 55 longitudinal OCT scans from 50 eyes of 50 individuals. DC (Dice Coefficient), Precision, and Recall of the proposed system are calculated. The suggested algorithm can achieve excellent segmentation accuracy without the need to segment the retinal layer. This approach has the potential to enhance clinical PED therapy and may be able to reliably segregate subretinal fluid from OCT images.