This project aims at developing and studying an optical imaging system in the presence of multiple scattering. The objective is to obtain three-dimensional quantitative images with a much better resolution than that of conventional microscopes. We consider a Digital Optical Tomographic Microscope whose principle consists in reconstructing numerically the permittivity distribution of the sample from the holograms obatined for different angles of illumination. In the presence of multiple scattering (between the object and the substrate or within the sample). In the presence of multiple scattering, the hologram of the object is not linked linearly to the permittivity distribution and depends on all the spatial frequencies of the sample. Thus, it is possible to obtain a much better resolution than that obtained under the assumption of linear imagery. In the first part, we will develop mathematical approaches using the Singular Value Decomposition and the theory of estimation for analyzing the quantity of information contained in the holograms. In a second part, we will implement these theoretical principles in an experimental set-up with calibrated samples. We will propose several inversion techniques for extracting iteratively the information provided by multiple scattering and we will evaluate their efficiency. The interest of adding a priori information in the inversion procedure will also be studied. The project concerns fundamental aspects on image formation in the presence of multiple scattering, numerical aspects on the development of reconstruction techniques, technological issues for making calibrated nano-objects, and experimental aspects for the development of a tomographic microscope set-up based on new concepts in optics. The partners of the project : Institut Fresnel, LPMA, L2S and LPN, whose domains of expertise are complementary, will study problems of physics with approaches that are original and new in the field of optical imagery.
