Self-Tuned Deep Super Resolution

Abstract

Deep learning has been successfully applied to image super resolution (SR). In this paper, we propose a deep joint super resolution (DJSR) model to exploit both external and self similarities for SR. A Stacked Denoising Convolutional Auto Encoder (SDCAE) is first pre-trained on external examples with proper data augmentations. It is then fine-tuned with multi-scale self examples from each input, where the reliability of self examples is explicitly taken into account. We also enhance the model performance by sub-model training and selection. The DJSR model is extensively evaluated and compared with state-of-the-arts, and show noticeable performance improvements both quantitatively and perceptually on a wide range of images.

Cited by in corpus (9)

• An All-in-One Network for Dehazing and Beyond
• Studying Very Low Resolution Recognition Using Deep Networks
• Image Super-Resolution via Dual-State Recurrent Networks
• Volumetric Super-Resolution of Multispectral Data
• Robust Emotion Recognition from Low Quality and Low Bit Rate Video: A Deep Learning Approach
• ${\mathbf{D}}^3$: Deep Dual-Domain Based Fast Restoration of JPEG-Compressed Images
• Self-supervised Fine-tuning for Correcting Super-Resolution Convolutional Neural Networks
• Designing A Composite Dictionary Adaptively From Joint Examples
• Orthogonally Regularized Deep Networks For Image Super-resolution