The Ninth NTIRE 2024 Efficient Super-Resolution Challenge Report

This paper provides a comprehensive review of the NTIRE 2024 challenge, focusing on efficient single-image super-resolution (ESR) solutions and their outcomes. The task of this challenge is to super-resolve an input image with a magnification factor of x4 based on pairs of low and corresponding high-resolution images. The primary objective is to develop networks that optimize various aspects such as runtime, parameters, and FLOPs, while still maintaining a peak signal-to-noise ratio (PSNR) of approximately 26.90 dB on the DIV2K_LSDIR_valid dataset and 26.99 dB on the DIV2K_LSDIR_test dataset. In addition, this challenge has 4 tracks including the main track (overall performance), sub-track 1 (runtime), sub-track 2 (FLOPs), and sub-track 3 (parameters). In the main track, all three metrics (ie runtime, FLOPs, and parameter count) were considered. The ranking of the main track is calculated based on a weighted sum-up of the scores of all other sub-tracks. In sub-track 1, the practical runtime performance of the submissions was evaluated, and the corresponding score was used to determine the ranking. In sub-track 2, the number of FLOPs was considered. The score calculated based on the corresponding FLOPs was used to determine the ranking. In sub-track 3, the number of parameters was considered. The score calculated based on the corresponding parameters was used to determine the ranking. RLFN is set as the baseline for efficiency measurement. The challenge had 262 registered participants, and 34 teams made valid submissions. They gauge the state-of-the-art in efficient single-image super-resolution. To facilitate the reproducibility of the challenge and enable other researchers to build upon these findings, the code and the pre-trained model of validated solutions are made publicly available at https://github.com/Amazingren/NTIRE2024_ESR/.

NTIRE 2024 Challenge on Image Super-Resolution ($\times$4): Methods and Results

This paper reviews the NTIRE 2024 challenge on image super-resolution ($\times$4), highlighting the solutions proposed and the outcomes obtained. The challenge involves generating corresponding high-resolution (HR) images, magnified by a factor of four, from low-resolution (LR) inputs using prior information. The LR images originate from bicubic downsampling degradation. The aim of the challenge is to obtain designs/solutions with the most advanced SR performance, with no constraints on computational resources (e.g., model size and FLOPs) or training data. The track of this challenge assesses performance with the PSNR metric on the DIV2K testing dataset. The competition attracted 199 registrants, with 20 teams submitting valid entries. This collective endeavour not only pushes the boundaries of performance in single-image SR but also offers a comprehensive overview of current trends in this field.

Learning Noise-Resistant Image Representation by Aligning Clean and Noisy Domains

Recent supervised and unsupervised image representation learning algorithms have achieved quantum leaps. However, these techniques do not account for representation resilience against noise in their design paradigms. Consequently, these effective methods suffer failure when confronted with noise outside the training distribution, such as complicated real-world noise that is usually opaque to model training. To address this issue, dual domains are optimized to separately model a canonical space for noisy representations, namely the Noise-Robust (NR) domain, and a twinned canonical clean space, namely the Noise-Free (NF) domain, by maximizing the interaction information between the representations. Given the dual canonical domains, we design a target-guided implicit neural mapping function to accurately translate the NR representations to the NF domain, yielding noise-resistant representations by eliminating noise regencies. The proposed method is a scalable module that can be readily integrated into existing learning systems to improve their robustness against noise. Comprehensive trials of various tasks using both synthetic and real-world noisy data demonstrate that the proposed Target-Guided Dual-Domain Translation (TDDT) method is able to achieve remarkable performance and robustness in the face of complex noisy images.

Parameter-Efficient Tuning Makes a Good Classification Head

In recent years, pretrained models revolutionized the paradigm of natural language understanding (NLU), where we append a randomly initialized classification head after the pretrained backbone, e.g. BERT, and finetune the whole model. As the pretrained backbone makes a major contribution to the improvement, we naturally expect a good pretrained classification head can also benefit the training. However, the final-layer output of the backbone, i.e. the input of the classification head, will change greatly during finetuning, making the usual head-only pretraining (LP-FT) ineffective. In this paper, we find that parameter-efficient tuning makes a good classification head, with which we can simply replace the randomly initialized heads for a stable performance gain. Our experiments demonstrate that the classification head jointly pretrained with parameter-efficient tuning consistently improves the performance on 9 tasks in GLUE and SuperGLUE.

Semantic-Aware Latent Space Exploration for Face Image Restoration

For image restoration, most existing deep learning based methods tend to overfit the training data leading to bad results when encountering unseen degradations out of the assumptions for training. To improve the robustness, generative adversarial network (GAN) prior based methods have been proposed, revealing a promising capability to restore photo-realistic and high-quality results. But these methods suffer from semantic confusion, especially on semantically significant images such as face images. In this paper, we propose a semantic-aware latent space exploration method for image restoration (SAIR). By explicitly modeling referenced semantics information, SAIR can consistently restore severely degraded images not only to high-resolution highly-realistic looks but also to correct semantics. Quantitative and qualitative experiments collectively demonstrate the effectiveness of the proposed SAIR. Our code can be found in https://github.com/Liamkuo/SAIR.

Functional Neural Networks for Parametric Image Restoration Problems

Almost every single image restoration problem has a closely related parameter, such as the scale factor in super-resolution, the noise level in image denoising, and the quality factor in JPEG deblocking. Although recent studies on image restoration problems have achieved great success due to the development of deep neural networks, they handle the parameter involved in an unsophisticated way. Most previous researchers either treat problems with different parameter levels as independent tasks, and train a specific model for each parameter level; or simply ignore the parameter, and train a single model for all parameter levels. The two popular approaches have their own shortcomings. The former is inefficient in computing and the latter is ineffective in performance. In this work, we propose a novel system called functional neural network (FuncNet) to solve a parametric image restoration problem with a single model. Unlike a plain neural network, the smallest conceptual element of our FuncNet is no longer a floating-point variable, but a function of the parameter of the problem. This feature makes it both efficient and effective for a parametric problem. We apply FuncNet to super-resolution, image denoising, and JPEG deblocking. The experimental results show the superiority of our FuncNet on all three parametric image restoration tasks over the state of the arts.

An Efficient Dual-reference Training Data Acquisition Method for CNN-Based Image Super-Resolution

For deep learning methods of image super-resolution, the most critical issue is whether the paired low and high resolution images for training accurately reflect the sampling process of real cameras. Low and high resolution (LR$\sim$HR) image pairs synthesized by existing degradation models (e.g. bicubic downsampling) deviate from those in reality; thus the super-resolution CNN trained by these synthesized LR$\sim$HR image pairs does not perform well when being applied to real images. In this paper, we propose a novel method to capture a large set of realistic LR$\sim$HR image pairs using real cameras. The data acquisition is carried out under controllable lab conditions with minimum human intervention and at high throughput (about 500 image pairs per hour). The high level of automation makes it easy to produce a set of real LR$\sim$HR training image pairs for each camera.Our innovation is to shoot images displayed on an ultra-high quality screen at different resolutions. There are three distinctive advantages of our method for image super-resolution. First, as the LR and HR images are taken of a 3D planar surface (the screen) the registration problem fits exactly to a homography model and we can display specially designed markers on the image to improve the registration precision. Second, the displayed digital image file can be exploited as a reference to optimize the high frequency content of the restored image. Third, this high-efficiency data collection method makes it possible to collect a customized dataset for each camera sensor, for which one can train a specific model for the intended camera sensor. Experimental results show that training a super-resolution CNN by our LR$\sim$HR dataset has superior restoration performance than training it by existing datasets on real world images at the inference stage.

An Efficient Dual-reference Training Data Acquisition Method for CNN Image Super-Resolution

For deep learning methods of image super-resolution, the most critical issue is whether the paired low and high resolution images for training accurately reflect the sampling process of real cameras. Low and high resolution (LR$\sim$HR) image pairs synthesized by existing degradation models (e.g. bicubic downsampling) deviate from those in reality; thus the super-resolution CNN trained by these synthesized LR$\sim$HR image pairs does not perform well when being applied to real images. In this paper, we propose a novel method to capture a large set of realistic LR$\sim$HR image pairs using real cameras. The data acquisition is carried out under controllable lab conditions with minimum human intervention and at high throughput (about 500 image pairs per hour). The high level of automation makes it easy to produce a set of real LR$\sim$HR training image pairs for each camera.Our innovation is to shoot images displayed on an ultra-high quality screen at different resolutions. There are three distinctive advantages of our method for image super-resolution. First, as the LR and HR images are taken of a 3D planar surface (the screen) the registration problem fits exactly to a homography model and we can display specially designed markers on the image to improve the registration precision. Second, the displayed digital image file can be exploited as a reference to optimize the high frequency content of the restored image. Third, this high-efficiency data collection method makes it possible to collect a customized dataset for each camera sensor, for which one can train a specific model for the intended camera sensor. Experimental results show that training a super-resolution CNN by our LR$\sim$HR dataset has superior restoration performance than training it by existing datasets on real world images at the inference stage.

Dual-reference Training Data Acquisition and CNN Construction for Image Super-Resolution

For deep learning methods of image super-resolution, the most critical issue is whether the paired low and high resolution images for training accurately reflect the sampling process of real cameras. Low and high resolution (LR$\sim$HR) image pairs synthesized by existing degradation models (\eg, bicubic downsampling) deviate from those in reality; thus the super-resolution CNN trained by these synthesized LR$\sim$HR image pairs does not perform well when being applied to real images. In this paper, we propose a novel method to capture a large set of realistic LR$\sim$HR image pairs using real cameras.The data acquisition is carried out under controllable lab conditions with minimum human intervention and at high throughput (about 500 image pairs per hour). The high level of automation makes it easy to produce a set of real LR$\sim$HR training image pairs for each camera. Our innovation is to shoot images displayed on an ultra-high quality screen at different resolutions.There are three distinctive advantages with our method that allow us to collect high-quality training datasets for image super-resolution. First, as the LR and HR images are taken of a 3D planar surface (the screen) the registration problem fits exactly to a homography model. Second, we can display special markers on the image margin to further improve the registration precision.Third, the displayed digital image file can be exploited as a reference to optimize the high frequency content of the restored image. Experimental results show that training a super-resolution CNN by our LR$\sim$HR dataset has superior restoration performance than training it by existing datasets on real world images at the inference stage.

Graph Based Link Prediction between Human Phenotypes and Genes

Background: The learning of genotype-phenotype associations and history of human disease by doing detailed and precise analysis of phenotypic abnormalities can be defined as deep phenotyping. To understand and detect this interaction between phenotype and genotype is a fundamental step when translating precision medicine to clinical practice. The recent advances in the field of machine learning is efficient to predict these interactions between abnormal human phenotypes and genes. Methods: In this study, we developed a framework to predict links between human phenotype ontology (HPO) and genes. The annotation data from the heterogeneous knowledge resources i.e., orphanet, is used to parse human phenotype-gene associations. To generate the embeddings for the nodes (HPO & genes), an algorithm called node2vec was used. It performs node sampling on this graph based on random walks, then learns features over these sampled nodes to generate embeddings. These embeddings were used to perform the downstream task to predict the presence of the link between these nodes using 5 different supervised machine learning algorithms. Results: The downstream link prediction task shows that the Gradient Boosting Decision Tree based model (LightGBM) achieved an optimal AUROC 0.904 and AUCPR 0.784. In addition, LightGBM achieved an optimal weighted F1 score of 0.87. Compared to the other 4 methods LightGBM is able to find more accurate interaction/link between human phenotype & gene pairs.