“…In general, the existing algorithms for data argumentation in through-the-wall radar imaging (TWRI) can be divided into five categories: signal domain transformation methods, statistical signal processing methods, compressed sensing (CS) methods, traditional machine learning methods and deep learning methods [3]. Among them, the signal domain transformation methods, including fast Fourier transform [4]- [7], shorttime Fourier transform [8], [9], wavelet transform [10], [11], and their improvements, [12]- [14] can achieve high peak signal-to-noise ratio (PSNR) and robustness with the lowest computational cost in specific conditions, which achieves the classifiers' accuracy up to 80% [6]- [8]. The statistical signal processing methods, including signal estimation algorithm [15]- [18] and signal detection algorithm [19]- [22], are applied with single-transmitter single-receiver (SISO), multitransmitter multiple-receiver (MIMO) and array through-thewall radar systems, which improve the recognition accuracy to more than 80% [15].…”