Deep Richardson-Lucy Deconvolution for Low-Light Image Deblurring
This addresses image quality issues for photographers and computer vision applications, but it is incremental as it builds on existing deblurring techniques with learned components.
The paper tackles the problem of deblurring low-light images with saturated pixels, which are challenging due to clipping effects, by proposing a data-driven method that models saturation with a learned latent map and integrates prior information into a Richardson-Lucy scheme, achieving favorable performance against state-of-the-art algorithms on synthetic and real-world images.
Images taken under the low-light condition often contain blur and saturated pixels at the same time. Deblurring images with saturated pixels is quite challenging. Because of the limited dynamic range, the saturated pixels are usually clipped in the imaging process and thus cannot be modeled by the linear blur model. Previous methods use manually designed smooth functions to approximate the clipping procedure. Their deblurring processes often require empirically defined parameters, which may not be the optimal choices for different images. In this paper, we develop a data-driven approach to model the saturated pixels by a learned latent map. Based on the new model, the non-blind deblurring task can be formulated into a maximum a posterior (MAP) problem, which can be effectively solved by iteratively computing the latent map and the latent image. Specifically, the latent map is computed by learning from a map estimation network (MEN), and the latent image estimation process is implemented by a Richardson-Lucy (RL)-based updating scheme. To estimate high-quality deblurred images without amplified artifacts, we develop a prior estimation network (PEN) to obtain prior information, which is further integrated into the RL scheme. Experimental results demonstrate that the proposed method performs favorably against state-of-the-art algorithms both quantitatively and qualitatively on synthetic and real-world images.