KBNet: Kernel Basis Network for Image Restoration

03/06/2023
by   Yi Zhang, et al.
0

How to aggregate spatial information plays an essential role in learning-based image restoration. Most existing CNN-based networks adopt static convolutional kernels to encode spatial information, which cannot aggregate spatial information adaptively. Recent transformer-based architectures achieve adaptive spatial aggregation. But they lack desirable inductive biases of convolutions and require heavy computational costs. In this paper, we propose a kernel basis attention (KBA) module, which introduces learnable kernel bases to model representative image patterns for spatial information aggregation. Different kernel bases are trained to model different local structures. At each spatial location, they are linearly and adaptively fused by predicted pixel-wise coefficients to obtain aggregation weights. Based on the KBA module, we further design a multi-axis feature fusion (MFF) block to encode and fuse channel-wise, spatial-invariant, and pixel-adaptive features for image restoration. Our model, named kernel basis network (KBNet), achieves state-of-the-art performances on more than ten benchmarks over image denoising, deraining, and deblurring tasks while requiring less computational cost than previous SOTA methods.

READ FULL TEXT

page 7

page 8

page 9

page 11

research
07/28/2021

Content-aware Directed Propagation Network with Pixel Adaptive Kernel Attention

Convolutional neural networks (CNNs) have been not only widespread but a...
research
09/13/2020

Attention Cube Network for Image Restoration

Recently, deep convolutional neural network (CNN) have been widely used ...
research
09/06/2022

CNSNet: A Cleanness-Navigated-Shadow Network for Shadow Removal

The key to shadow removal is recovering the contents of the shadow regio...
research
06/19/2020

Concatenated Attention Neural Network for Image Restoration

In this paper, we present a general framework for low-level vision tasks...
research
08/19/2021

Real-time Image Enhancer via Learnable Spatial-aware 3D Lookup Tables

Recently, deep learning-based image enhancement algorithms achieved stat...
research
09/23/2022

Modular Degradation Simulation and Restoration for Under-Display Camera

Under-display camera (UDC) provides an elegant solution for full-screen ...
research
07/22/2020

Learning Disentangled Feature Representation for Hybrid-distorted Image Restoration

Hybrid-distorted image restoration (HD-IR) is dedicated to restore real ...

Please sign up or login with your details

Forgot password? Click here to reset