Abstract:As a fundamental task in imaging and computer vision, image denoising has remained an active topic of discussion over recent decades. Traditionally, a denoiser is designed by inferring analytical priors in the image,1,2 whose accuracy is restrained by its specified form. Recently, the rapid growth in computational hardware has reignited research on neural networks. Image denoising has also been widely investigated in relation to neural networks. Given sufficient image samples and advanced training techniques,3,4 convolutional neural networks (CNNs)5,6 have achieved more promising results than prior-based methods. However, current CNN denoisers remain far from achieving a realistic noise-free performance. First, real-image noise consists of complex noise components;7 therefore, it is highly Deep convolutional neural networks (CNNs) have achieved considerable success in terms of image denoising. However, previous CNN denoisers have been restricted by rigid kernel convolution that applies equal spatial treatment across images. To fully utilize the local differences, we propose a kernel prediction network that examines each pixel region and predicts unique pixel-wise kernels. Several optimizations have been further designed to gather sufficient information for single-image denoising task. We adopt dilated residual blocks to view the local pixel region at varying receptive fields. Then, kernel fusion assembles the information from different scopes and generates accurate kernels for each pixel. Instead of applying the predicted kernels to the original image, we construct a compressed feature map as a substitution such that more relevant local features are collected. Experiments are used to demonstrate that our network achieves favorable results compared with state-of-the-art methods and is adequate for practical applications. (c) 2021 SPIE and IS&T

ADCNN: Towards learning adaptive dilation for convolutional neural networks

Spatial-Adaptive Network for Single Image Denoising

Dilated kernel prediction network for

ASCNet: Adaptive-Scale Convolutional Neural Networks for Multi-Scale Feature Learning

Adaptive Fractional Dilated Convolution Network for Image Aesthetics Assessment

Generalized Dilation Neural Networks

Adaptive deformable convolutional network

Frequency-Adaptive Dilated Convolution for Semantic Segmentation

Dilated Residual Encode-Decode Networks for Image Denoising

Content-aware convolutional neural networks

DDCNet: Deep Dilated Convolutional Neural Network for Dense Prediction

Densely connected multidilated convolutional networks for dense prediction tasks

Adaptive Gabor convolutional networks

A Multiscale Image Denoising Algorithm Based On Dilated Residual Convolution Network

DTS: Dynamic Training Slimming with Feature Sparsity for Efficient Convolutional Neural Network

Dilated Convolution with Learnable Spacings makes visual models more aligned with humans: a Grad-CAM study

Adaptive Distributed Convolutional Neural Network Inference at the Network Edge with ADCNN

Adaptive Pixel-wise Structured Sparse Network for Efficient CNNs

TAda! Temporally-Adaptive Convolutions for Video Understanding

EADNet: Efficient Asymmetric Dilated Network for Semantic Segmentation

Dilated Convolution with Learnable Spacings: beyond bilinear interpolation