Revisiting Global Statistics Aggregation for Improving Image Restoration

by   Xiaojie Chu, et al.

Global spatial statistics, which are aggregated along entire spatial dimensions, are widely used in top-performance image restorers. For example, mean, variance in Instance Normalization (IN) which is adopted by HINet, and global average pooling (i.e. mean) in Squeeze and Excitation (SE) which is applied to MPRNet. This paper first shows that statistics aggregated on the patches-based/entire-image-based feature in the training/testing phase respectively may distribute very differently and lead to performance degradation in image restorers. It has been widely overlooked by previous works. To solve this issue, we propose a simple approach, Test-time Local Statistics Converter (TLSC), that replaces the region of statistics aggregation operation from global to local, only in the test time. Without retraining or finetuning, our approach significantly improves the image restorer's performance. In particular, by extending SE with TLSC to the state-of-the-art models, MPRNet boost by 0.65 dB in PSNR on GoPro dataset, achieves 33.31 dB, exceeds the previous best result 0.6 dB. In addition, we simply apply TLSC to the high-level vision task, i.e. semantic segmentation, and achieves competitive results. Extensive quantity and quality experiments are conducted to demonstrate TLSC solves the issue with marginal costs while significant gain. The code is available at


page 1

page 3

page 12

page 13

page 14

page 15

page 16

page 17


HINet: Half Instance Normalization Network for Image Restoration

In this paper, we explore the role of Instance Normalization in low-leve...

DualNorm-UNet: Incorporating Global and Local Statistics for Robust Medical Image Segmentation

Batch Normalization (BN) is one of the key components for accelerating n...

Simple Baselines for Image Restoration

Although there have been significant advances in the field of image rest...

ParseNet: Looking Wider to See Better

We present a technique for adding global context to deep convolutional n...

HighEr-Resolution Network for Image Demosaicing and Enhancing

Neural-networks based image restoration methods tend to use low-resoluti...

FedDisco: Federated Learning with Discrepancy-Aware Collaboration

This work considers the category distribution heterogeneity in federated...

Image Restoration with Mean-Reverting Stochastic Differential Equations

This paper presents a stochastic differential equation (SDE) approach fo...

Please sign up or login with your details

Forgot password? Click here to reset