Title :
A perturbation analysis of the regularized constrained total least squares
Author :
Fan, X. ; Younan, N.H. ; Taylor, C.D.
Author_Institution :
Tomographic Imaging Lab., Cincinnati Univ., OH, USA
fDate :
2/1/1996 12:00:00 AM
Abstract :
A perturbation analysis of the regularized constrained total least squares technique (RCTLS) is performed to ascertain the statistical characteristics of its solution in signal restoration. With an appropriate choice of a regularization parameter, it is shown that the RCTLS method reduces the mean square error (MSE) and variance estimates. Also, a MSE comparison with traditional techniques for different signal-to-noise ratios (SNR) is made to verify its applicability
Keywords :
least squares approximations; perturbation techniques; signal restoration; statistics; RCTLS method; SNR; mean square error; perturbation analysis; regularization parameter; regularized constrained total least squares; signal restoration; signal-to-noise ratios; statistical characteristics; variance estimates; Inverse problems; Least squares methods; Maximum likelihood estimation; Mean square error methods; Performance analysis; Semiconductor device noise; Signal analysis; Signal restoration; Signal to noise ratio; Sparse matrices;
Journal_Title :
Circuits and Systems II: Analog and Digital Signal Processing, IEEE Transactions on
