Precise characterization of concealed person-worn objects will speed up the passenger screening process by reducing the rate of nuisance alarms, while also enhancing the airport security imaging systems. This paper presents an automatic, real-time method for wideband millimeter-wave radar identification of the nominal surface contours of the human body – even with affixed foreign objects or when a segment of the body cross-section is not captured by the radar – without relying on the body's bilateral symmetry. The developed method is verified experimentally when applied to the actual images generated by a laboratory airport scanning prototype developed recently by the US Department of Homeland Security (DHS). Our method uses the noisy collection of radar cross-section reflectivity data to extract the main contours and estimates the nominal body surface cross-sections through fitting a small-term Fourier series of circumferential variation. This is a necessary step for accurate characterizing of concealed terrorist threat objects affixed to the body.

We highlight the need for – and describe initial strategies to find – new digital-optical basis functions and performance merit functions to serve as a foundation for designing, analyzing, characterizing, testing, and comparing a range of computational imaging systems. Such functions will provide a firm theoretical foundation for computational sensing and imaging and enhanced design software, thereby broadly speeding the development of computational imaging systems.