$\mathbf{D^3}$: Deep Dual-Domain Based Fast Restoration of JPEG-Compressed Images
Preprint
Abstract
In this paper, we design a Deep Dual-Domain ($\mathbf{D^3}$) based fast restoration model to remove artifacts of JPEG compressed images. It leverages the large learning capacity of deep networks, as well as the problem-specific expertise that was hardly incorporated in the past design of deep architectures. For the latter, we take into consideration both the prior knowledge of the JPEG compression scheme, and the successful practice of the sparsity-based dual-domain approach. We further design the One-Step Sparse Inference (1-SI) module, as an efficient and light-weighted feed-forward approximation of sparse coding. Extensive experiments verify the superiority of the proposed $D^3$ model over several state-of-the-art methods. Specifically, our best model is capable of outperforming the latest deep model for around 1 dB in PSNR, and is 30 times faster.
Keywords
All Related Versions
This publication has 23 references indexed in Scilit:
- Compression Artifacts Reduction by a Deep Convolutional Network2015
- Deep Networks for Image Super-Resolution with Sparse Prior2015
- Convolutional neural networks at constrained time cost2015
- Data-driven sparsity-based restoration of JPEG-compressed images in dual transform-pixel domain2015
- Learning a Deep Convolutional Network for Image Super-Resolution2014
- A Learning-Based Approach to Reduce JPEG Artifacts in Image Matting2013
- Image deblocking via sparse representationSignal Processing: Image Communication, 2012
- Loss-Specific Training of Non-Parametric Image Restoration Models: A New State of the Art2012
- Iterative Thresholding for Sparse ApproximationsJournal of Fourier Analysis and Applications, 2008
- Real-time compression artifact reduction via robust nonlinear filtering1999