A Cosine Network for Image Super-Resolution

2025-12-23

IEEE Transactions on Image Processing

AI Summary

This work addresses the insufficient exploitation of structural information in image super-resolution by proposing CSRNet, a novel architecture that employs heterogeneous even-odd modules to extract complementary homologous structural features. By integrating both linear and nonlinear structural cues, the model enhances the robustness of its learned representations. To further improve optimization, cosine annealing learning rate scheduling is introduced during training, effectively mitigating convergence to suboptimal local minima. Despite its architectural simplicity, CSRNet achieves significant improvements in reconstruction quality, delivering performance on par with state-of-the-art methods across multiple benchmark datasets.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Structured Learning
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Bridging structured and unstructured data

Abstract

Deep convolutional neural networks can use hierarchical information to progressively extract structural information to recover high-quality images. However, preserving the effectiveness of the obtained structural information is important in image super-resolution. In this paper, we propose a cosine network for image super-resolution (CSRNet) by improving a network architecture and optimizing the training strategy. To extract complementary homologous structural information, odd and even heterogeneous blocks are designed to enlarge the architectural differences and improve the performance of image super-resolution. Combining linear and non-linear structural information can overcome the drawback of homologous information and enhance the robustness of the obtained structural information in image super-resolution. Taking into account the local minimum of gradient descent, a cosine annealing mechanism is used to optimize the training procedure by performing warm restarts and adjusting the learning rate. Experimental results illustrate that the proposed CSRNet is competitive with state-of-the-art methods in image super-resolution.

Problem

• image super-resolution
• structural information
• deep convolutional neural networks
• high-quality image reconstruction

Innovation

• cosine annealing
• heterogeneous blocks
• structural information
• image super-resolution
• deep convolutional neural networks