Tumor response extraction based on ensemble empirical mode decomposition for early breast cancer detection by UWB

Ultra wide-band (UWB) microwave imaging technique is very promising in detecting the early breast cancer. In previous studies, the tumor-free model was always needed to help extract the calibration waveform, which cannot be obtained directly in the real clinical detection. In this paper, to solve this problem, a novel signal processing method based on ensemble empirical mode decomposition (EEMD) is proposed to extract the tumor response. This approach is crucial to extract the tumor response directly from the detected signals in the actual clinical diagnosis, as it does not need the tumor-free model at all. The feasibility of using this method to extract the tumor response signal without the tumor-free model is explained in the reconstructed breast image, which shows the correct tumor information.