Adaptive optics imaging of retinal microstructures: Image processing for medical applications

Pâques, Michel; Rossant, Florence; Lermé, Nicolas; Miloudi, Chahira; Kulcsár, Caroline; Sahel, José‐Alain; Loquin, Kevin; Mugnier, Laurent M.; Bloch, Isabelle; Koch, Edouard

doi:10.1109/iwcim.2014.7008813

Cited by 2 publications

(1 citation statement)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As recently reported, flood illumination adaptive optics (FIAO) imaging allows for sensitive detection of retinal vasculitis [2]. FIAO is a recent technology that enables the visualization of retinal microstructures such as photoreceptors [3] and arterial walls [4], noninvasively [5]. The resolution of the Rtx1 camera [6] used in our study is about 1.6µm/pixel.…”

Section: Introductionmentioning

confidence: 77%

Segmentation of retinal vessels in adaptive optics images for assessment of vasculitis

Lagarrigue-Charbonnier

Rossant

Bloch

et al. 2016

2016 Sixth International Conference on Image Processing Theory, Tools and Applications (IPTA)

Self Cite

View full text Add to dashboard Cite

International audience—In this paper we propose a new method for segmenting retinal vessels in adaptive optics images. This method is particularly dedicated for segmenting vessels with significant morphological alterations due to vasculitis, but it is also accurate for vessels with moderate or without alteration. It relies on a pre-segmentation step which is crucial for the robustness and accuracy of the results. This step is based on a specific morphological processing of isolines of the original image: they constitute of good basis for the segmentation because they are disposed along the wall borders of the vessels. Regularization is then performed using active contour model embedding a parallelism constraint. This novel model allows precise segmenting inner and outer walls of the vessel. In particular it is more accurate in the case of vasculitis than the existing methods. This is the only method that allows quantification. The results and the runtime make it suitable for clinical use

show abstract