Closed-Loop Matters: Dual Regression Networks for Single Image Super-Resolution

By Yong Guo, Jian Chen, Jingdong Wang, Qi Chen, Jiezhang Cao, Zeshuai Deng, Yanwu Xu, Mingkui Tan. 2020

319 citations recorded by the source corpus.

Deep neural networks have exhibited promising performance in image super-resolution (SR) by learning a nonlinear mapping function from low-resolution (LR) images to high-resolution (HR) images. However, there are two underlying limitations to existing SR methods. First, learning the mapping function from LR to HR images is typically an ill-posed problem, because there exist infinite HR images that can be downsampled to the same LR image. As a result, the space of the possible functions can be extremely large, which makes it hard to find a good solution. Second, the paired LR-HR data may be unavailable in real-world applications and the underlying degradation method is often unknown. For such a more general case, existing SR models often incur the adaptation problem and yield poor performance. To address the above issues, we propose a dual regression scheme by introducing an additional constraint on LR data to reduce the space of the possible functions. Specifically, besides the mapping from LR to HR images, we learn an additional dual regression mapping estimates the down-sampling kernel and reconstruct LR images, which forms a closed-loop to provide additional supervision. More critically, since the dual regression process does not depend on HR images, we can directly learn from LR images. In this sense, we can easily adapt SR models to real-world data, e.g., raw video frames from YouTube. Extensive experiments with paired training data and unpaired real-world data demonstrate our superiority over existing methods.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Radu Timofte
3. Kai Zhang
4. Jian Chen
5. Mingkui Tan
6. Jingyun Liang
7. Ming-Hsuan Yang
8. Lei Zhang

Citations in the corpus

• Flow-based Kernel Prior with Application to Blind Super-Resolution (2021, background): “With the development of convolutional neural networks (CNN), CNN-based methods [10, 19, 25, 30, 38, 47, 56] have been gaining the popularity in solving image SR.”
• Residual Feature Distillation Network for Lightweight Image Super-Resolution (2020, background): “[6] developed a dual regression scheme by introducing an additional constraint such that the mappings can form a closed-loop and LR images can be reconstructed to enhance the performance of SR models.”
• VRT: A Video Restoration Transformer (2022, methodology): “Similar to image restoration [18], [19], [20], learning-based methods, especially CNN-based methods, have become the primary workhorse for video restoration [2], [5], [12], [21], [22], [23], [24], [25], [26], [27] in recent years.”
• Breaking the Curse of Space Explosion: Towards Efficient NAS with Curriculum Search (2020, methodology): “…stateof-the-art results in many challenging tasks including image classification (Krizhevsky et al., 2012; Liu et al., 2020; Zheng et al., 2015; Guo et al., 2020a), semantic segmentation (Shelhamer et al., 2017; Chen et al., 2018; Huang et al., 2019), and many other areas (Zheng et al.,…”