Abstract
Numerous single image blind deblurring algorithmshave been proposed to restore latent sharp images under camera motion. However, these algorithms are mainly eval-uated using either synthetic datasets or few selected realblurred images. It is thus unclear how these algorithmswould perform on images acquired “in the wild” and how we could gauge the progress in the field. In this paper, weaim to bridge this gap. We present the first comprehensiveperceptual study and analysis of single image blind deblur-ring using real-world blurred images. First, we collect adataset of real blurred images and a dataset of synthetically blurred images. Using these datasets, we conduct a large-scale user study to quantify the performance of several representative state-of-the-art blind deblurring algorithms. Second, we systematically analyze subject preferences, including the level of agreement, significance tests of score differences, and rationales for preferring one methodover another. Third, we study the correlation between hu-man subjective scores and several full-reference and noreference image quality metrics. Our evaluation and analysis indicate the performance gap between synthetically blurred images and real blurred image and sheds light on future research in single image blind deblurring.