基于L^1拟合与光滑正则化的图像去噪声问题能够转化为一个非光滑方程.在此基础上,证明了非光滑方程是强半光滑的,因而解此方程的广义牛顿法具有局部二次收敛性.
Image denosing problem based on L^1-fitting and smooth regularization can be reformulated as a system of nonsmooth equations. On the basis of this reformulation, it is proved that the system of nonsmooth equations is strongly semismooth so that the generalized Newton method for solving this system possesses locally quadratic convergence.