Medical Imaging

Multi Angle OCT (Optical Coherence Tomography)

The software presented in this video results from our work conducted since 2009 in collaboration with Telecom ParisTech and the CIC of the Quinze-Vingts hospital in the domain of OCT imaging.

This software, dedicated to the segmentation of multi-angle OCT images, has been developped at ISEP with the financial support of Satt Lutech.

M. Pâques, F. Rossant, S. Mrejen, A. Pedinielli, Characterization of the anisotropy of photoreceptors and RPE by multiangle OCT, ARVO, Denver, Colorado, May 2015.

Parallel Snakes

The proposed model, developped jointly by ISEP (F. Rossant) and Telecom ParisTech (I. Bloch), in collaboration with N. Lermé and I. Ghorbel, has been applied on several kinds of images:

- OCT images, for the segmentation of retinal layers [Pattern Recognition 2015] [ICIP 2011]
- Standard eye fundus images, for the segmentation of vessel branches
- Adaptive Optics images, for the segmentation of retinal artery wall [Pattern Recognition Letters-2016] [ICPR 2015][ICIAR 2015] and bifucations [ICIP-19]
- CNN-based segmentation of retinal vessels in adaptive optics images [IPTA-20]

NB: select the best quality on your You Tube player or upload high-quality mp4 (bottom of this page)

Segmentation of arterial walls in adaptive optics images of the retina

Segmentation of a retinal interface, based on a fixed reference line:

Segmentation of the hyper reflective complex, based on 2 curves moving under approximate parallelism constraint:

Segmentation of an arterial bifurcations in Adaptive Optics.