Dry Noncontact Microwave Breast Imaging Based on Variational Mode Decomposition

A method based on variational mode decomposition (VMD) is proposed in this article for suppressing skin artifacts to enable early detection of small and weakly scattering breast lesions using dry and noncontact microwave imaging. In contrast to conventional approaches that rely on the coupling medium or precalibrated waveforms, the proposed algorithm integrates VMD with a singular-value curvature spectrum (SVCS) to adaptively determine the number of modes and employs$K$-means clustering assisted by the whale optimization algorithm (WOA) to separate skin responses from tissue responses in multichannel acquisitions. Experiments on an MRI-derived, 3-D-printed breast phantom demonstrate effective suppression of strong skin backscatter and accurate isolation of breast tissue responses in the absence of the coupling medium. Comparative analyses with established methods, including ensemble empirical mode decomposition (EEMD) and singular value decomposition (SVD), indicate that the proposed approach mitigates mode mixing and truncation order uncertainty and yields higher detection quality and improved localization while providing superior adaptability and a low computational cost compatible with near real-time clinical application. These attributes make it a more practical and efficient candidate for clinical microwave breast imaging (MBI).