Alexandre Aubry

Scattering reflection matrix approach to ultra-deep imaging through biological media (Conference Presentation)

When imaging with classical waves, multiple scattering (MS) is often seen as an unavoidable obstacle. The diffraction-limited resolution obtainable with methods such as microscopy requires that single-scattering (SS) dominates; for depths where MS processes become important, such methods result in an image without any connection to the reflectivity of the medium. Conversely, techniques such as diffuse optical tomography take advantage of the diffuse nature of light, but their resolution power is limited. To do better, methods such as wavefront shaping and adaptive optics have been developed. Focussing through a thick diffusive layer was demonstrated using a transmission matrix approach consisting of the measurement of Green’s functions between each pixel of a spatial light modulator (SLM) and of a charge-coupled device (CCD) camera across the medium. To image inside a multiple-scattering medium, we present a matrix approach based on the experimental measurement of a reflection matrix from the medium. An analysis based on the geometric and statistical properties of this reflection matrix can enhance the SS contribution which would otherwise be swamped by MS at large depths, and correct the resulting image for aberration effects induced by the turbid medium itself. The correction does not require the presence of bright scatterers, does not rely on any feedback loop and works even at depths where the field-of-view contains several isoplanatic patches. Here we present the application of our reflection matrix approach to optical imaging in biological tissues. Compared to OCT and related methods, we demonstrate an extension of the current imaging-depth limit.

Multiple scattering filter: Application to plane defect detection in a nickel alloy

The ultrasonic inspection of polycrystalline media remains a challenge. The high noise levels due to interaction between the wave and the microstructure limit the efficiency of classical ultrasonic techniques to detect a defect in a coarse grain structure. The aim of this work is to reduce the influence of multiple scattering in order to increase the information obtained from the defect. The technique introduced here is based on array probes for the acquisition of the medium’s response matrix by full matrix capture, after which a filter based on random matrix theory is applied. Here an improvement of this technique is applied on nickel-based alloy mock-ups that present an unfavourable grain structure and well known bulk and plane defects. The results in normal incidence and with an angle array probe of 128 elements and 5u2005MHz of central frequency are compared to classical phased array probe techniques.