Coherent sub-space processing for time reversal microwave imaging

We consider electromagnetic inverse scattering using time reversal for microwave imaging for early stage breast cancer detection. Assuming that the malignant tissue inside the breast to be the strongest scatterer, we localize it by employing subspace based time reversal techniques viz., decomposition of the time reversal operator (DORT) and time reversal MUSIC. In denser breasts, the fibroglandular tissue clutter dominates the back scattered signals at the receiver. Also, the received scattered field becomes highly correlated thereby significantly affecting the performance of subspace based time reversal algorithms. To alleviate this, we employ coherent signal subspace transformation to obtain a coherently focussed time reversal operator (TRO). We test the performance of proposed method using MRI derived realistic numerical breast phantom of a moderately dense breast.