Research‎ > ‎

PhD Thesis Proposal 2

PhD in medical image processing and analysis: Modeling of the lamina cribrosa of the retina from 3D OCT data

  • Grade : M2, with a specialty in image processing and analysis
  • Doctoral school  EDITE  : 
  • Supervisors: Prof. Florence Rossant (ISEP) ; Prof. Isabelle Bloch (Télécom ParisTech), Prof. Michel Pâques (CIC de l’hôpital des Quinze-Vingts)
    • Télécom ParisTech - Dept TSI - 46 rue Barrault - 75013 Paris
    • ISEP, Dept. SITe, 10 rue de Vanves 92130, Issy les Moulineaux
    • CIC, Hôpital des Quinze-Vingts, 28 rue de Charenton, 75012 PARIS
  • Starting date: last trimestre of 2017
  • Required skills: image processing and analysis, segmentation, active contour models, mathematical morphology, filtering,  image registration, neural networks, Matlab and C/C++ programming.
  • Other skills: good level in english, team work.
  • Application: apply on EDITE's website from  21-04-2017 to  19-05-2017  to compete for EDITE's PhD grants  - subject on EDITE's website


Glaucoma is one of the leading causes of blindness in the world. Although its physiopathology remains unclear, a major role of the deformations of the lamina cribrosa (LC), the structure located at the interface between the ocular globe and the optic nerve, is strongly suspected [1,2]. The LC is a three-dimensional porous structure composed of flexible collagenous tissue through which all the nerve fibers from the retina pass to join the brain. During glaucoma, it is likely that the deformations of the LC induce a deformation of the pores and hence damage bypassing neurons.

Optical Coherence Tomography (OCT) is a scanning ocular imaging technique that uses infrared light to observe the fine structures of the retina and the optic nerve. It offers the possibility of reconstructing in three dimensions the structures that are imaged. For this reason, we wish to study the lamina cribrosa with this imaging modality, and in particular the 3D arrangement of the LC pores revealing the pathways of the axons. Additionally, adaptive optics (AO) imaging allows observing the pores of LC in en-faces images with a higher resolution, close to the micron.

So the aim of this PhD subject is to propose methods for the segmentation and modeling of axonal pathways from 3D SD-OCT (FIG. 1) and AO data, which has never been done before. Methods will be developed to register AO and OCT images and initialize the detection of LC pores from AO data. Then, the LC pores will be tracked in the 3D OCT volume, to achieve a 3D modeling of the axonal pathways (FIG. 2). Combining several 3D OCT images acquired at different orientations will be investigated, in order to improve the signal to noise ratio and obtain more accurate and reliable segmentations. This will imply to develop registration procedures, which will have to take into account local distortions due to the scanning process inherent to OCT technology. From the segmentation results, we will compute novel biomarkers to provide a morphometric analysis of the LC. This will allow for the first time the observation of the possible causes of neuronal loss during glaucoma. Longitudinal follow-up will be a key element in determining disease progression.

This project will be conducted in close partnership with the Clinical Investigation Center of the Quinze-Vingts Hospital (Paris). It is part of a larger project aiming at modeling the anatomy and blood flow of the optical nerve head from several imaging modalities. The partners involved in the supervision of this PhD thesis have collaborated for more than ten years on projects related to the study of the retina from several imaging modalities.


FIG. 2. 3D SD-OCT data displayed in 3 orthogonal planes