Image Restoration Using Convolutional Auto-encoders with Symmetric Skip Connections

29 Jun 2016Xiao-Jiao MaoChunhua ShenYu-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... (read more)

PDF Abstract

Evaluation results from the paper


Task Dataset Model Metric name Metric value Global rank Compare
Image Denoising BSD200 sigma10 RED30 PSNR 33.63 # 1
Image Denoising BSD200 sigma10 RED30 SSIM 0.9319 # 1
Image Denoising BSD200 sigma30 RED30 PSNR 27.95 # 1
Image Denoising BSD200 sigma30 RED30 SSIM 0.8019 # 1
Image Denoising BSD200 sigma50 RED30 PSNR 25.75 # 1
Image Denoising BSD200 sigma50 RED30 SSIM 0.7167 # 1
Image Denoising BSD200 sigma70 RED30 PSNR 24.37 # 1
Image Denoising BSD200 sigma70 RED30 SSIM 0.6551 # 1
Image Denoising Urban100 sigma50 RED30 PSNR 26.32 # 5
Image Denoising Urban100 sigma70 RED30 PSNR 24.63 # 3