2018 PIRM Challenge on Perceptual Image Super-resolution

By Yochai Blau, Roey Mechrez, Radu Timofte, Tomer Michaeli, Lihi Zelnik-Manor. 2018

545 citations recorded by the source corpus.

This paper reports on the 2018 PIRM challenge on perceptual super-resolution (SR), held in conjunction with the Perceptual Image Restoration and Manipulation (PIRM) workshop at ECCV 2018. In contrast to previous SR challenges, our evaluation methodology jointly quantifies accuracy and perceptual quality, therefore enabling perceptual-driven methods to compete alongside algorithms that target PSNR maximization. Twenty-one participating teams introduced algorithms which well-improved upon the existing state-of-the-art methods in perceptual SR, as confirmed by a human opinion study. We also analyze popular image quality measures and draw conclusions regarding which of them correlates best with human opinion scores. We conclude with an analysis of the current trends in perceptual SR, as reflected from the leading submissions.

Subfields: Computer Vision and Pattern Recognition, Evaluation & Benchmarks

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. Lei Zhang
4. Chen Change Loy
5. Ming-Hsuan Yang
6. Chao Dong
7. Dacheng Tao
8. Kai Zhang

In Evaluation & Benchmarks

1. Chao Dong
2. Chen Change Loy
3. Kai Zhang
4. Steven C. H. Hoi
5. Lei Zhang
6. Luc Van Gool
7. Yu Qiao
8. Shuhang Gu

Citations in the corpus

• Designing a Practical Degradation Model for Deep Blind Image Super-Resolution (2021, methodology): “Since the ground-truth for the RealSRSet dataset is not available, we adopt the non-reference image quality assessment (IQA) metrics including NIQE [38], NRQM [32] and PI [4] for quantitative evaluation.”
• ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks (2018, methodology): “We take a variant of ESRGAN to participate in the PIRM-SR Challenge [5].”
• PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models (2020, methodology): “To provide another measure of perceptual quality, we evaluated the Naturalness Image Quality Evaluator (NIQE) score [16], previously used in perceptual super-resolution [10, 4, 24].”
• SinGAN: Learning a Generative Model From a Single Natural Image (2019, methodology): “Following [4], we compare these 5 methods in Table 3 on the BSD100 dataset [35] in terms of distortion (RMSE) and perceptual quality (NIQE [40]), which are two fundamentally conflicting requirements [5].”
• TOPIQ: A Top-Down Approach From Semantics to Distortions for Image Quality Assessment (2023, methodology): “Traditional approaches for NR-IQA rely on natural scene statistics (NSS) [3], [26]–[30].”
• Structure-Preserving Super Resolution With Gradient Guidance (2020, methodology): “We choose Perceptual Index (PI) [6], Learned Perceptual Image Patch Similarity (LPIPS) [50], PSNR and Structure Similarity (SSIM) [43] as the evaluation metrics.”
• Residual Non-local Attention Networks for Image Restoration (2019, background): “Recently, great progresses have been made in image restoration community, where Timofte et al. (Timofte et al., 2017), Ancuti et al. (Ancuti et al., 2018), and Blau et al. (Blau et al., 2018) lead the main competitions recently and achieved new research status and records.”
• Enhanced Pix2pix Dehazing Network (2019, background): “And it has been recognized to be effective in image super-resolution [2].”
• Towards Real-World Blind Face Restoration with Generative Facial Prior (2021, background): “Note that the pixelwise metrics PSNR and SSIM are not correlation well with the subjective evaluation of human observers [2, 43] and our model is not good at these two metrics.”
• Deep Learning for Image Super-Resolution: A Survey (2019, methodology): “PIRM [47] 200 (617, 482) 292, 021 PNG environments, flora, natural scenery, objects, people, etc.”
• A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications (2020, methodology): “Benefiting from these improvements, ESRGAN won the first place in the PIRM2018-SR Challenge (region 3) [316] and got the best perceptual index.”
• Analyzing Perception-Distortion Tradeoff using Enhanced Perceptual Super-resolution Network (2018, methodology): “(b) Performance comparison of top 9 methods from PIRM-SR challenge [4].”
• AIM 2019 Challenge on Constrained Super-Resolution: Methods and Results (2019, background): “In PIRM 2018, the SR challenges focused on improving the perceptual quality of super-resolved images [2] and efficiency on smartphones [10].”
• Toward Real-World Single Image Super-Resolution: A New Benchmark and a New Model (2019, background): “SISR has been an active research topic for decades [39, 59, 46, 48, 4, 6] because of its high practical values in enhancing image details and textures.”
• The Unreasonable Effectiveness of Texture Transfer for Single Image Super-resolution (2018, background): “ceptual image enhancement challenge (PIRM) [31] made a huge step to promote perceptual enhancment in images.”
• Residual Dense Network for Image Restoration (2018, methodology): “Among them, deep learning (DL)-based methods have achieved dramatic advantages against conventional methods in computer vision [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46].”
• Low-Light Image Enhancement with Wavelet-Based Diffusion Models (2023, methodology): “In row 3, Zero-DCE, EnlightenGAN, and DDIM smooth the cloud texture, as Uformer and WeatherDiff produce gridding effects and messy content.”