Defocused Image Restoration Using Translation Invariant Wavelet Transform

In this paper, we restore a blurred image caused by defocus of a lens using the wavelet transform. In a defocus blurred image, the blurring kernel varies depending on the position in image, so the positional frequency representation such as the wavelet space is necessary for deblurring them. For both removing noise and restoring decreased higher frequency components, we use the translation invariant wavelet transform realized by the RI-spline wavelets. Experimental results using synthesized images, under the assumption that the blurring kernel of any position in a image is known, show that the high denoising performance of the translation invariant wavelet shrinkage enables better deblurring performance