Estimating reflectivity by non-stationary deconvolution of ultra-wide band radar data: Application to microwave breast imaging

The quality of a microwave breast image is gaged in part by the resolution of the recorded scatter fields. Deconvolution is applied to improve the signal resolution and yield a representation of subsurface reflectivity by compressing the excitation pulse in the recorded signals. As the electromagnetic wave propagates through lossy and dispersive breast tissues, the waveform is distorted due to the frequency dependent attenuation. This imposes challenges on the stationary deconvolution techniques, which assume a time-invariant propagating wavelet. In this paper, we apply Gabor non-stationary deconvolution to compensate for these dissipative and dispersive effects. Results demonstrate that this method provides a more accurate estimate of the reflectivity of the imaged object and shows ability to compensate for the wave attenuation and correct the waveform distortion.