UWB microwave imaging for breast tumor detection based on shrinkage covariance matrix estimation

This paper presents a new covariance matrix estimate based on shrinkage method for breast tumor detection application. The parameters of shrinkage covariance matrix are obtained by solving a modified semi-definite programming (SDP) convex problem based on the mean-squared error (MSE) criterion. The reconstructed covariance matrix, using the determined shrinkage parameters, is then used to replace the conventional sample covariance matrix (SCM) in the beamformer. The performance of the resulting beamformer is tested by a 2-D numerical breast analysis model and the simulations show the proposed approach possesses a better target identification capability and improves the signal-to-clutter-noise ratio (SCNR).