Convolutional Sparse Modular Fusion Algorithm for Non-Rigid Registration of Visible–Infrared Images

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(5), P. 2508 - 2508

Published: Feb. 26, 2025

Existing image fusion algorithms involve extensive models and high computational demands when processing source images that require non-rigid registration, which may not align with the practical needs of engineering applications. To tackle this challenge, study proposes a comprehensive framework for convolutional sparse in context registration visible–infrared images. Our approach begins an attention-based encoder to extract cross-modal feature encodings from enhance extraction, we introduce feature-guided loss information entropy guide extraction homogeneous isolated features, resulting decomposition network. Next, create module estimates parameters based on pairs. Finally, develop by applying filtering groups, high-quality fused maximized retention. Experimental results multiple datasets indicate that, compared similar studies, proposed algorithm achieves average improvement 8.3% 30.6% performance mutual information. In addition, downstream target recognition tasks, generated show maximum 20.1% relative accuracy original Importantly, our maintains relatively lightweight parameter load.

Language: Английский

Denoising diffusion fusion network for semantic segmentation based on degradation analysis modeling with graph networks

Information Fusion, Journal Year: 2025, Volume and Issue: unknown, P. 103205 - 103205

Published: April 1, 2025

Language: Английский