A multiple-parameter generalization of the Tikhonov-Miller regularization method for image restoration

The Tikhonov-Miller regularization approach has long been utilized for restoring images that are contaminated by noise and are blurred due for example to camera defocusing or linear motion. It is posed as a least squares approximation problem in the l₂ space that provides a parameterized tradeoff between accuracy and smoothness of the restored image, with the tradeoff being controlled by a scalar regularization parameter. The authors generalize the Tikhonov-Miller method by incorporating multiple regularization parameters into the regularization process. As the regularization parameters can be chosen based on a frequency-domain criterion to better balance the approximation accuracy and solution smoothness, the proposed method leads to improved restoration of degraded noisy images as compared to the conventional regularization method