Burst image super-resolution (BISR) aims to enhance the resolution of a keyframe by leveraging information from multiple low-resolution images captured in quick succession. In the deep learning era, BISR methods have evolved from fully convolutional networks to transformer-based architectures, which, despite their effectiveness, suffer from the quadratic complexity of self-attention. We see Mamba as the next natural step in the evolution of this field, offering a comparable global receptive field and selective information routing with only linear time complexity. In this work, we introduce BurstMamba, a Mamba-based architecture for BISR. Our approach decouples the task into two specialized branches: a spatial module for keyframe super-resolution and a temporal module for subpixel prior extraction, striking a balance between computational efficiency and burst information integration. To further enhance burst processing with Mamba, we propose two novel strategies: (i) optical flow-based serialization, which aligns burst sequences only during state updates to preserve subpixel details, and (ii) a wavelet-based reparameterization of the state-space update rules, prioritizing high-frequency features for improved burst-to-keyframe information passing. Our framework achieves SOTA performance on public benchmarks of SyntheticSR, RealBSR-RGB, and RealBSR-RAW.

翻译：连拍图像超分辨率（BISR）旨在通过利用快速连续拍摄的多张低分辨率图像中的信息，来提升关键帧的分辨率。在深度学习时代，BISR方法已从全卷积网络演进至基于Transformer的架构；尽管这些架构有效，但其自注意力机制具有二次复杂度。我们认为Mamba是该领域发展的下一个自然阶段，它在仅具有线性时间复杂度的同时，提供了可比的全局感受野和选择性信息路由能力。本工作中，我们提出了BurstMamba，一种基于Mamba的BISR架构。我们的方法将任务解耦为两个专用分支：用于关键帧超分辨率的空间模块，以及用于亚像素先验提取的时间模块，从而在计算效率与连拍信息整合之间取得平衡。为了进一步利用Mamba增强连拍处理，我们提出了两种新颖策略：（i）基于光流的序列化方法，仅在状态更新时对齐连拍序列以保留亚像素细节；（ii）基于小波的状态空间更新规则重参数化，优先处理高频特征以改善从连拍到关键帧的信息传递。我们的框架在SyntheticSR、RealBSR-RGB和RealBSR-RAW公开基准测试中达到了最先进的性能。