IRSRMamba: Infrared Image Super-Resolution via Mamba-based Wavelet Transform Feature Modulation Model

2024-05-16

arXiv.org

AI Summary

Infrared image super-resolution suffers from detail loss and structural incoherence due to uniform backgrounds, weak edges, and sparse textures. To address this, we propose Mamba-WaveNet—the first framework integrating the Mamba state-space model with discrete wavelet transform (DWT), enabling linear-complexity long-range dependency modeling and adaptive multi-scale feature modulation. We design an end-to-end differentiable wavelet reconstruction pipeline and introduce a self-supervised semantic consistency loss to mitigate contextual discontinuities induced by patch-based processing. Extensive experiments on multiple infrared benchmark datasets demonstrate that our method achieves significant PSNR and SSIM improvements over state-of-the-art approaches, accompanied by superior visual quality. The source code is publicly available.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Infrared image super-resolution demands long-range dependency modeling and multi-scale feature extraction to address challenges such as homogeneous backgrounds, weak edges, and sparse textures. While Mamba-based state-space models (SSMs) excel in global dependency modeling with linear complexity, their block-wise processing disrupts spatial consistency, limiting their effectiveness for IR image reconstruction. We propose IRSRMamba, a novel framework integrating wavelet transform feature modulation for multi-scale adaptation and an SSMs-based semantic consistency loss to restore fragmented contextual information. This design enhances global-local feature fusion, structural coherence, and fine-detail preservation while mitigating block-induced artifacts. Experiments on benchmark datasets demonstrate that IRSRMamba outperforms state-of-the-art methods in PSNR, SSIM, and perceptual quality. This work establishes Mamba-based architectures as a promising direction for high-fidelity IR image enhancement. Code are available at https://github.com/yongsongH/IRSRMamba.

Problem

• Enhances infrared image super-resolution quality
• Addresses long-range dependency and multi-scale feature extraction
• Mitigates block-induced artifacts in image reconstruction

Innovation

• Wavelet transform feature modulation
• Mamba-based semantic consistency loss
• Enhanced global-local feature fusion