Dense Cross-Scale Image Alignment With Fully Spatial Correlation and Just Noticeable Difference Guidance

2025-11-12

AI Summary

To address the low alignment accuracy and high computational cost in unsupervised image alignment, this paper proposes the Dense Cross-Scale Alignment Model (DCAM). Methodologically: (1) a full-spatial correlation module is designed to model pixel-level long-range dependencies, enhancing robustness to misalignment; (2) a Just-Noticeable Difference (JND)-aware mechanism is introduced to guide the model toward human-perceptually salient distortion regions; and (3) a cross-scale feature fusion architecture is developed to enable flexible trade-offs between accuracy and efficiency. Extensive experiments on multiple benchmark datasets demonstrate that DCAM consistently outperforms state-of-the-art methods, achieving simultaneous improvements in alignment accuracy, visual quality, and inference speed. These results validate both the effectiveness and practicality of the proposed approach.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Security and Privacy: Large-scale security measurements

Abstract

Existing unsupervised image alignment methods exhibit limited accuracy and high computational complexity. To address these challenges, we propose a dense cross-scale image alignment model. It takes into account the correlations between cross-scale features to decrease the alignment difficulty. Our model supports flexible trade-offs between accuracy and efficiency by adjusting the number of scales utilized. Additionally, we introduce a fully spatial correlation module to further improve accuracy while maintaining low computational costs. We incorporate the just noticeable difference to encourage our model to focus on image regions more sensitive to distortions, eliminating noticeable alignment errors. Extensive quantitative and qualitative experiments demonstrate that our method surpasses state-of-the-art approaches.

Problem

• Unsupervised image alignment has limited accuracy and high complexity
• Model improves alignment using cross-scale feature correlations
• Incorporates spatial correlation and perceptual guidance to reduce errors

Innovation

• Dense cross-scale image alignment model
• Fully spatial correlation module improves accuracy
• Just noticeable difference guides distortion focus