Statistical Modeling of Retinal Optical Coherent Tomography


In this project, another model for retinal Optical Coherence Tomography (OCT) pictures is proposed. This measurable model depends on acquainting a nonlinear Gaussianization change with believer the likelihood dissemination work (pdf) of every OCT intra-retinal layer to a Gaussian appropriation. The retina is a layered structure and in OCT every one of these layers has a particular pdf which is adulterated by spot commotion, in this manner a blend display for measurable demonstrating of OCT pictures is proposed. A Normal-Laplace conveyance, which is a convolution of a Laplace pdf and Gaussian clamor, is proposed as the circulation of every segment of this model.

The purpose behind picking Laplace pdf is the monotonically rotting conduct of OCT powers in each layer for sound cases. Subsequent to fitting a blend model to the information, every segment is gaussianized and every one of them are joined by Averaged Maximum A Posterior (AMAP) technique. To show the capacity of this technique, another difference upgrade strategy in view of this measurable model is proposed and tried on thirteen sound 3D OCTs taken by the Topcon 3D OCT and five 3D OCTs from Age-related Macular Degeneration (AMD) patients, taken by Zeiss Cirrus HD-OCT. Contrasting the outcomes and two battling strategies, the noticeable quality of the proposed technique is exhibited both outwardly and numerically. Moreover, to demonstrate the viability of the proposed technique for a more straightforward and particular reason, a change in the division of intra-retinal layers utilizing the proposed differentiate improvement strategy as a preprocessing step, is illustrated.

BASE PAPER: Statistical Modeling of Retinal Optical Coherent Tomography


Please enter your comment!
Please enter your name here