Image Restoration Using Deep Regulated Convolutional Networks

19 Oct 2019  ·  Peng Liu, Xiaoxiao Zhou, Junyi Yang, El Basha Mohammad D, Ruogu Fang ·

While the depth of convolutional neural networks has attracted substantial attention in the deep learning research, the width of these networks has recently received greater interest. The width of networks, defined as the size of the receptive fields and the density of the channels, has demonstrated crucial importance in low-level vision tasks such as image denoising and restoration. However, the limited generalization ability, due to the increased width of networks, creates a bottleneck in designing wider networks. In this paper, we propose the Deep Regulated Convolutional Network (RC-Net), a deep network composed of regulated sub-network blocks cascaded by skip-connections, to overcome this bottleneck. Specifically, the Regulated Convolution block (RC-block), featured by a combination of large and small convolution filters, balances the effectiveness of prominent feature extraction and the generalization ability of the network. RC-Nets have several compelling advantages: they embrace diversified features through large-small filter combinations, alleviate the hazy boundary and blurred details in image denoising and super-resolution problems, and stabilize the learning process. Our proposed RC-Nets outperform state-of-the-art approaches with significant performance gains in various image restoration tasks while demonstrating promising generalization ability. The code is available at

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Image Super-Resolution BSD100 - 2x upscaling RC-Net PSNR 31.86 # 20
Image Super-Resolution BSD100 - 3x upscaling RC-Net PSNR 28.76 # 16
Image Super-Resolution BSD100 - 4x upscaling RC-Net PSNR 27.21 # 41
Grayscale Image Denoising BSD200 sigma10 RC-Net PSNR 36.36 # 1
Grayscale Image Denoising BSD200 sigma30 RC-Net PSNR 33.57 # 1
Grayscale Image Denoising BSD200 sigma50 RC-Net PSNR 32.48 # 1
Grayscale Image Denoising BSD200 sigma70 RC-Net PSNR 31.17 # 1
Image Super-Resolution Set5 - 2x upscaling RC-Net PSNR 37.42 # 25
Image Super-Resolution Set5 - 3x upscaling RC-Net PSNR 33.43 # 19
Image Super-Resolution Set5 - 4x upscaling RC-Net PSNR 31.01 # 50