Two real-time reconstruction algorithms, quantitative microwave holography and scattered-power mapping, have been shown to be successful in the imaging of compressed tissue of relatively small thicknesses such as 1 cm and 2 cm. In both cases, planar data acquisition of frequency-swept transmission coefficients has been employed. Despite the fact that these algorithms are based on a linear forward model of scattering, they have been capable of providing quantitative estimates of the tissue permittivity due to the experimentally derived kernel of the scattering integral. Here, we demonstrate similar performance with a thicker (about 5 cm) compressed-breast phantom. This thickness is greater or comparable to the median thickness employed in mammography, depending on the view (craniocaudal or mediolateral oblique). The two methods are described in a common mathematical framework for the first time. The importance of the system calibration and the choice of a host medium are discussed through experiments. A new method for focusing onto suspect regions is demonstrated. The limitations of real-time imaging are highlighted along with an outlook to improving the image resolution and suppressing artifacts without sacrificing the reconstruction speed.

An ultra-wide band (UWB) active slot antenna for tissue sensing arrays is developed and measured to operate from 3 GHz to 8 GHz. The sensing element reported here is to be used in the imaging of the breast. It integrates a low-noise amplifier (LNA) with a printed-slot antenna to achieve gain enhancement of about 20 dB. The first prototype integrates the biasing of the LNA on the same board as the antenna. The second prototype uses an external bias-tee providing DC power to the LNA through the coaxial connector of the active antenna. Both prototypes achieve the expected gain enhancement in comparison with the passive antenna element.

A prototype of a bias-switched active sensor was developed and measured to establish the achievable dynamic range in a new generation of active arrays for microwave tissue imaging. The sensor integrates a printed slot antenna, a low-noise amplifier (LNA) and an active mixer in a single unit, which is sufficiently small to enable inter-sensor separation distance as small as 12 mm. The sensor’s input covers the bandwidth from 3 GHz to $7.5$ GHz. Its output intermediate frequency (IF) is 30 MHz. The sensor is controlled by a simple bias-switching circuit, which switches ON and OFF the bias of the LNA and the mixer simultaneously. It was demonstrated experimentally that the dynamic range of the sensor, as determined by its ON and OFF states, is 109 dB and 118 dB at resolution bandwidths of 1 kHz and 100 Hz, respectively.

Recently, a real-time deconvolution-based reconstruction method has been proposed that performs successful quantitative imaging of complex tissue objects of thicknesses around 1 cm. Here, imaging of 5-cm thick phantoms is demonstrated experimentally. The microwave system aims at breast imaging using planar acquisition similar to that in mammography. An important conclusion has been reached that Rytov's approximation in the forward model, as compared to Born's approximation, may lead to superior ability to localize scatterers amidst clutter and noise. Practical aspects of the acquisition system are also discussed including the choice of the medium used as padding around the breast phantom.