Shrinkage Fields for Effective Image Restoration

Abstract

Many state-of-the-art image restoration approaches do not scale well to larger images, such as megapixel images common in the consumer segment. Computationally expensive optimization is often the culprit. While efficient alternatives exist, they have not reached the same level of image quality. The goal of this paper is to develop an effective approach to image restoration that offers both computational efficiency and high restoration quality. To that end we propose shrinkage fields, a random field-based architecture that combines the image model and the optimization algorithm in a single unit. The underlying shrinkage operation bears connections to wavelet approaches, but is used here in a random field context. Computational efficiency is achieved by construction through the use of convolution and DFT as the core components; high restoration quality is attained through loss-based training of all model parameters and the use of a cascade architecture. Unlike heavily engineered solutions, our learning approach can be adapted easily to different trade-offs between efficiency and image quality. We demonstrate state-of-the-art restoration results with high levels of computational efficiency, and significant speedup potential through inherent parallelism.

Comparison of Denoising Algorithms for Urban Scenes

“…In the shrinkage fields approach, a restored image is predicted by finding the MAP estimate of the image given the degraded image where the corruption process is modeled with a Gaussian likelihood kernel and a strength term (Schmidt and Roth, 2014). A block coordinate descent strategy is used that alternates between minimizing with respect to in ideal image x and auxiliary variables within z.…”

“…The latest denoising methods have demonstrated impressive results, however, their levels of effectiveness seem to differ by only small amounts, so it is not entirely clear which method works best for any particular application or which method should be selected. (Buades et al, 2005;Chatterjee and Milanfar, 2009;Chen et al, 2013;Dabov et al, 2007;Dong et al, 2015;Gu et al, 2014;Schmidt and Roth, 2014). Further research into this area is extensive.…”

“…A third impressive new method uses shrinkage fields that are based on existing optimization algorithms for common random field models and are computationally efficient (Schmidt and Roth, 2014). This approach attains its performance through loss-based training of all model parameters and the use of a cascade architecture that can be adapted to different trade-offs between efficiency and image quality.…”

Fast approach to infrared image restoration based on shrinkage functions calibration

“…However, the mid-frequency textures are removed, because the independence among local gradients fails to capture the global statistics of gradients for the whole imag. 19 The learning-based method 8,14,15 capturing the prior from training data of sharp images can be used for restoration, because common patterns behind natural images can generate all kinds of natural structures. 20 Restored results of learning-based method are quite impressive, but the speed is too slow to restore in real time.…”

“…Therefore, how to model prior information from typical images to greatly improve image restoration becomes the most interesting area of image processing. [8][9][10][11][12][13][14][15] Generally, the prior is captured from natural images, which are sharp photographs of the typical environment in which we live. It can be grouped into two categories: heavy-tailed distribution and learning-based methods.…”