Image Super-Resolution via Progressive Cascading Residual Network

By Namhyuk Ahn, Byungkon Kang, Kyung-ah Sohn. 2018

71 citations recorded by the source corpus.

The problem of enhancing the resolution of a single low-resolution image has been popularly addressed by recent deep learning techniques. However, many deep learning approaches still fail to deal with extreme super-resolution scenarios because of the instability of training. In this paper, we address this issue by adapting a progressive learning scheme to the deep convolutional neural network. In detail, the overall training proceeds in multiple stages so that the model gradually increases the output image resolution. In our experiments, we show that this property yields a large performance gain compared to the non-progressive learning 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

Citations in the corpus

Image Super-Resolution via Iterative Refinement (2021, background): “sion based and trained with an MSE loss [49], [52], [54], [55], [56].”
SRFlow: Learning the Super-Resolution Space with Normalizing Flow (2020, methodology): “ethods do not reproduce the sharp details present in natural images due to their reliance on L 2 and L 1 reconstruction losses. This was addressed in URDGN [51], SRGAN [24] and more recent approaches [36,2,17,44] by adopting a conditional GAN based architecture and training strategy. While these wor”
NTIRE 2018 Challenge on Single Image Super-Resolution: Methods and Results (2018, methodology): “In the training process, extra CARN module is added after at the end of the stage and replace the previous reconstruction layers with the one that produces the image in double resolution.”
Fine-Grained Visual Classification via Progressive Multi-Granularity Training of Jigsaw Patches (2020, background): “cover large-scale structure of the image distribution and then shift attention to increasingly ner scale details. Recently, progressive training strategy has been widely utilized for generation tasks [19,27,32,1], since it can simplify the information propagation within the network by intermediate s”
Multi-stage image denoising with the wavelet transform (2022, background): “[1], we can see that stacked convolutional layers can refine obtained features.”
SinSR: Diffusion-Based Image Super-Resolution in a Single Step (2023, unknown): “One prevalent approach of early works is to train a regression model using paired training data [1, 2, 16, 44].”
Unsupervised Learning for Real-World Super-Resolution (2019, background): “While traditionally addressed with classical techniques [19, 12, 29, 36, 15], current approaches [9, 10, 22, 23, 26, 11, 2, 3, 14, 16] employ deep learning methodologies to train a mapping from LR to HR.”
Deep Learning for Image Super-Resolution: A Survey (2019, background): “[150] propose ProSR, ADRSR and progressive CARN, respectively, which are progressive not only on architectures (Sec.”
NTIRE 2020 Challenge on Real-World Image Super-Resolution: Methods and Results (2020, background): “search topic for decades [33, 17, 56, 65, 63, 69, 70, 71, 61, 13, 28, 64, 14, 15, 38, 40, 43, 16, 2, 3, 26, 29, 21] due to its many applications.”