Evaluation of image deblurring algorithms for real-time applications

Camera shake is a well-known source of degradation in digital images, as it introduces motion blur. Taking satisfactory photos under dim lighting conditions or using a hand-held camera is challenging. Same problems arise when camera is connected to mechanical equipments, that transfer vibrations to the camera itself. Since decades, many different theories and algorithms have been proposed with the aim of retrieving latent images from blurry inputs; most of them work quite well, but very often incur in large execution times. There are cases in which images have to be analyzed looking for features to be extracted; in this cases, it may be useful to consider deblurring as a preprocessing stage, that should not affect the performances of the whole image processing architecture, in terms of throughput. In this paper, an extensive survey of the deblurring algorithms that have been developed during the last 40 years is provided. Aim of this paper is to highlight software approaches that are able to quickly process input images and obtain good quality outcomes, analyzing the possibility of an hardware implementation to meet real-time requirements.