Simple Baselines for Image Restoration

TL;DR

This paper introduces NAFNet, a simple, nonlinear activation-free image restoration model that outperforms state-of-the-art methods in efficiency and accuracy across multiple benchmarks.

Contribution

The paper presents a novel, simplified network architecture for image restoration that removes nonlinear activations, achieving better performance with less computational cost.

Findings

NAFNet achieves 33.69 dB PSNR on GoPro, surpassing previous SOTA by 0.38 dB.

NAFNet achieves 40.30 dB PSNR on SIDD, exceeding previous SOTA by 0.28 dB.

NAFNet is more computationally efficient, using only 8.4% and less than half of the computational costs for respective tasks.

Abstract

Although there have been significant advances in the field of image restoration recently, the system complexity of the state-of-the-art (SOTA) methods is increasing as well, which may hinder the convenient analysis and comparison of methods. In this paper, we propose a simple baseline that exceeds the SOTA methods and is computationally efficient. To further simplify the baseline, we reveal that the nonlinear activation functions, e.g. Sigmoid, ReLU, GELU, Softmax, etc. are not necessary: they could be replaced by multiplication or removed. Thus, we derive a Nonlinear Activation Free Network, namely NAFNet, from the baseline. SOTA results are achieved on various challenging benchmarks, e.g. 33.69 dB PSNR on GoPro (for image deblurring), exceeding the previous SOTA 0.38 dB with only 8.4% of its computational costs; 40.30 dB PSNR on SIDD (for image denoising), exceeding the previous SOTA 0.28 dB with less than half of its computational costs. The code and the pre-trained models are released at https://github.com/megvii-research/NAFNet.