Ricerc@Sapienza

The point-spread function of an optical system imaged by a camera carries, in its geometrical details, key information on the optical components that are part of the system. In this work we explore the possibility of detecting anomalous structures buried under the skin surface by studying the deformations of the ideal Airy-disk point-spread function of a terahertz microscope, which are observed when the skin becomes one of the reflecting surfaces of the optical system. A custom terahertz microscope working in reflection mode at an angle of incidence of 45 degree has been built. A continuous-wave amplier-multiplier chain with output at 0.611 THz (~ 491 micron ) is used as the point source. A microbolometer camera optimized for THz detection is used for imaging the Airy disk. Skin simulants based on collagen gel with and without buried structures have been analyzed. Airy disks with and without deformations have been observed and geometrical features are automatically extracted from the point-spread function images. A machine learning approach based on the decision-tree algorithm produces signicant classication of unknown samples, based on the geometrical features of the point-spread function.