Ph. D. Project
Title:
Multimodal spectro-imaging and machine learning methods for bio-optical characterization of skin cancers
Dates:
2024/10/01 - 2027/09/30
Supervisor(s):
Other supervisor(s):
Prof. Hervé RINNERT (herve.rinnert@univ-lorraine.fr)
Description:
The thesis is part of the SpectroSkin project, winner of the CNRS 80PRIME 2024 action, for the creation of a new scientific collaboration between the CRAN and IJL teams, by exploiting their complementary skills in characterization optics of materials and skin. The challenge of the SpectroSkin project is, through a multi-scale interdisciplinary approach and by coupling for the first time complementary information from four experimental methodologies (diffuse reflectance, autofluorescence, Raman and Far Infrared), to make the link between biophotonic characteristics ( spectral signatures, optical parameters) of healthy and pathological skin samples (i) at the macroscopic scale in vivo, on the one hand and (ii) at the microscopic scale ex vivo, on the other hand. Machine learning (automatic classification) techniques aided by light-tissue interaction modeling methods will be developed to analyze the original multidimensional datasets thus generated and identify new combinations of discriminative bio-optical markers to more effectively solve the problem. in vivo diagnosis of skin cancers.
Keywords:
Tissue optics, skin, imaging, hyperspectral, modelling, machine learning
Conditions:
Duration : 36 months
Emloyer : CRAN UMR 7039 UL-CNRS (Université de Lorraine)
Places : CRAN and IJL, in Vandoeuvre-Lès-Nancy
Emloyer : CRAN UMR 7039 UL-CNRS (Université de Lorraine)
Places : CRAN and IJL, in Vandoeuvre-Lès-Nancy
Department(s):
| Biology, Signals and Systems in Cancer and Neuroscience |
Funds:
CNRS 80PRIME 2024
Publications: