Image Restoration Using Convolutional Auto-encoders with Symmetric Skip Connections

29 Jun 2016  ·  Xiao-Jiao Mao, Chunhua Shen, Yu-Bin Yang ·

Image restoration, including image denoising, super resolution, inpainting, and so on, is a well-studied problem in computer vision and image processing, as well as a test bed for low-level image modeling algorithms. In this work, we propose a very deep fully convolutional auto-encoder network for image restoration, which is a encoding-decoding framework with symmetric convolutional-deconvolutional layers... In other words, the network is composed of multiple layers of convolution and de-convolution operators, learning end-to-end mappings from corrupted images to the original ones. The convolutional layers capture the abstraction of image contents while eliminating corruptions. Deconvolutional layers have the capability to upsample the feature maps and recover the image details. To deal with the problem that deeper networks tend to be more difficult to train, we propose to symmetrically link convolutional and deconvolutional layers with skip-layer connections, with which the training converges much faster and attains better results. read more

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Image Super-Resolution BSD100 - 2x upscaling RED30 PSNR 31.99 # 10
SSIM 0.8974 # 5
Image Super-Resolution BSD100 - 3x upscaling RED30 PSNR 28.93 # 6
SSIM 0.7994 # 4
Image Super-Resolution BSD100 - 4x upscaling RED30 PSNR 27.4 # 31
SSIM 0.729 # 32
Grayscale Image Denoising BSD200 sigma10 RED30 PSNR 33.63 # 2
SSIM 0.9319 # 1
Grayscale Image Denoising BSD200 sigma30 RED30 PSNR 27.95 # 3
SSIM 0.8019 # 1
Grayscale Image Denoising BSD200 sigma50 RED30 PSNR 25.75 # 3
SSIM 0.7167 # 1
Grayscale Image Denoising BSD200 sigma70 RED30 PSNR 24.37 # 3
SSIM 0.6551 # 1
JPEG Artifact Correction Live1 (Quality 10 Grayscale) RED30 PSNR 29.35 # 8
JPEG Artifact Correction LIVE1 (Quality 20 Grayscale) RED30 PSNR 31.73 # 7
Image Super-Resolution Set14 - 2x upscaling RED30 PSNR 32.94 # 17
SSIM 0.9144 # 9
Image Super-Resolution Set14 - 3x upscaling RED30 PSNR 29.61 # 8
SSIM 0.8341 # 3
Image Super-Resolution Set14 - 4x upscaling RED30 PSNR 27.86 # 41
SSIM 0.7718 # 37
Image Super-Resolution Set5 - 2x upscaling RED30 PSNR 37.66 # 14
SSIM 0.9599 # 9
Image Super-Resolution Set5 - 3x upscaling RED30 PSNR 33.82 # 11
SSIM 0.923 # 6
Image Super-Resolution Set5 - 4x upscaling RED30 PSNR 31.51 # 37
SSIM 0.8869 # 36

Methods