Deep Gaussian Conditional Random Field Network: A Model-based Deep Network for Discriminative Denoising
This addresses image denoising for computer vision applications, offering a novel architecture but is incremental in combining existing concepts.
The paper tackles image denoising by proposing a deep Gaussian Conditional Random Field network that models input noise variance to handle multiple noise levels, outperforming state-of-the-art methods on Berkeley segmentation and PASCALVOC datasets.
We propose a novel deep network architecture for image\\ denoising based on a Gaussian Conditional Random Field (GCRF) model. In contrast to the existing discriminative denoising methods that train a separate model for each noise level, the proposed deep network explicitly models the input noise variance and hence is capable of handling a range of noise levels. Our deep network, which we refer to as deep GCRF network, consists of two sub-networks: (i) a parameter generation network that generates the pairwise potential parameters based on the noisy input image, and (ii) an inference network whose layers perform the computations involved in an iterative GCRF inference procedure.\ We train the entire deep GCRF network (both parameter generation and inference networks) discriminatively in an end-to-end fashion by maximizing the peak signal-to-noise ratio measure. Experiments on Berkeley segmentation and PASCALVOC datasets show that the proposed deep GCRF network outperforms state-of-the-art image denoising approaches for several noise levels.