Research

Retinal Vessel Extraction from 3D OCT

A semi-automated segmentation of slice-by-slice OCT stacks devices was successfully realized.

November 2011 - December 2013
Cyrill Gyger

Based on the findings from the master thesis “Object Segmentation in Speckle Noise”, this research project was launched in collaboration with Peter Maloca, MD. The goal of the project is to develop an automated mechanism to extract retinal blood vessels from OCT volume scans captured by commercial OCT devices. Beside image denoising and segmentation, image registration and other processing steps are subject to this project, required by the often insufficient output of commercial OCT scanners. So far, a semi-automated segmentation of slice-by-slice OCT stacks captured by commercial OCT devices (880 nm) was successfully realized.