Accelerating GMM-based patch priors for image restoration: Three ingredients for a 100$\times$ speed-up

TL;DR

The paper introduces FEPLL, a significantly faster version of the EPLL image restoration algorithm that maintains high quality, enabling real-time processing for various inverse problems without specialized hardware.

Contribution

It proposes three approximations to EPLL, creating FEPLL, which achieves over 100x speed-up with minimal quality loss, applicable to multiple image restoration tasks.

Findings

FEPLL is over 100 times faster than EPLL.

Maintains image quality with less than 0.5 dB loss.

Restores 512x512 images in under 0.5 seconds.

Abstract

Image restoration methods aim to recover the underlying clean image from corrupted observations. The Expected Patch Log-likelihood (EPLL) algorithm is a powerful image restoration method that uses a Gaussian mixture model (GMM) prior on the patches of natural images. Although it is very effective for restoring images, its high runtime complexity makes EPLL ill-suited for most practical applications. In this paper, we propose three approximations to the original EPLL algorithm. The resulting algorithm, which we call the fast-EPLL (FEPLL), attains a dramatic speed-up of two orders of magnitude over EPLL while incurring a negligible drop in the restored image quality (less than 0.5 dB). We demonstrate the efficacy and versatility of our algorithm on a number of inverse problems such as denoising, deblurring, super-resolution, inpainting and devignetting. To the best of our knowledge, FEPLL is the first algorithm that can competitively restore a 512x512 pixel image in under 0.5s for all the degradations mentioned above without specialized code optimizations such as CPU parallelization or GPU implementation.