SCALE ADAPTIVE BLIND DEBLURRING

摘要

Techniques are disclosed for removing blur from a single image by accumulating a blur kernel estimation across several scale levels of the image and balancing the contributions of the different scales to the estimation depending on the noise level in each observation. In particular, a set of observations can be obtained by applying a set of variable scale filters to a single blurry image at different scale levels. A single blur kernel can be estimated across all scales from the set of observations and used to obtain a single latent sharp image. The estimation at a large scale level is refined using the observations at successively smaller scale levels. The filtered observations may be weighted during the estimation to balance the contributions of each scale to the estimation of the blur kernel. A deblurred digital image is recovered by deconvolving the blurry digital image using the estimated blur kernel.

主权项

1. A computer-implemented digital image processing method comprising:
receiving, by a processor, input data representing a first observation of a blurry digital image filtered at a first scale level, and a second observation of the blurry digital image filtered a second scale level that is smaller than the first scale level; calculating a first estimated blur kernel based on the first observation; calculating a second estimated blur kernel based on the second observation using the first estimated blur kernel as a baseline; iteratively accumulating an estimated blur kernel from the first estimated blur kernel and the second estimated blur kernel; and deconvolving, by the processor, the blurry digital image using the estimated blur kernel to obtain a deblurred digital image.