Hybrid Clutter Rejection Technique for Improved Microwave Head Imaging

The accuracy of microwave head imaging is adversely affected by strong clutters that can completely mask the target response. To that end, different clutter removal techniques are modified for multistatic frequency-based imaging. It is shown that some deficiencies of those methods in time domain, such as time overlapping, can be alleviated when they are modified for use in frequency domain. Based on the explored performance of different methods in the frequency domain, a hybrid technique, which combines the benefits of average subtraction and entropy-based filtering methods, is proposed. In this method, the average value of the multistatic scattered signals is subtracted from them at each frequency sample to remove late-stage clutters, whereas an entropy-based method is applied to mitigate early-stage strong clutters. The proposed technique is verified in realistic environments using simulations and experiments. The utilized system for verification is 1.1-3.2 GHz frequency-domain multistatic with an eight-element antenna array, and compact microwave transceiver. The simulations are performed on MRI-derived head model, whereas the experiments are done on realistic artificial head phantom. The obtained results from different locations and sizes of emulated brain injuries confirm the effectiveness of the proposed method in producing high quality images of the head after mitigating the clutter.