Estimating image blur in the wavelet domain.

In this paper, a wavelet based method is proposed to estimate the blur in an image using information contained in the image itself. We look at the sharpness of the sharpest edges in the blurred image, which contain information about the blurring. Specifically, a smoothness measure, the Lipschitz exponent, is computed for these sharpest edges. Its value is related to the variance of a gaussian point spread function. This value is only dependent on the blur in the image and not on the image contents. This allows us to estimate the variance of the blur directly from the image itself. With minor modifications, the method can be extended to other types of blur that are described by one parameter (airy disks, out-of-focus blur, …).