用Mamba架构实现遥感大模型自监督预训练,突破图像分辨率与规模瓶颈。
RoMA: Scaling up Mamba-based Foundation Models for Remote Sensing
- 针对遥感图像设计旋转感知预训练与多尺度词元预测机制。
- 在多个任务上超越ViT模型,精度更高且计算更高效。
- 适合遥感视觉、大规模自监督学习研究者参考。
视觉变换器(ViTs)的自监督学习进展推动了遥感(RS)基础模型的发展,但自注意力的二次复杂度严重限制了其可扩展性,尤其在大模型和高分辨率图像上。虽然线性复杂度的Mamba架构提供了潜在替代方案,但现有遥感应用仍局限于小规模、特定领域的有监督任务。为此,我们提出RoMA框架,利用大规模、多样化的未标注数据,实现基于Mamba的遥感基础模型的可扩展自监督预训练。RoMA通过定制的自回归学习策略提升高分辨率图像的可扩展性,包含两项关键创新:1)结合自适应裁剪与角度嵌入的旋转感知预训练机制,以处理稀疏分布且方向任意的目标;2)多尺度词元预测目标,应对遥感图像中极端的对象尺度变化。系统性实证研究验证了Mamba遵循遥感数据与参数缩放定律,性能随模型和数据规模增长而稳定提升。跨场景分类、目标检测和语义分割任务的实验表明,RoMA预训练的Mamba模型在准确率和计算效率上均持续优于基于ViT的模型。代码与预训练模型将公开于 https://github.com/MiliLab/RoMA。
原文摘要 · Abstract (English)
Recent advances in self-supervised learning for Vision Transformers (ViTs) have fueled breakthroughs in remote sensing (RS) foundation models. However, the quadratic complexity of self-attention poses a significant barrier to scalability, particularly for large models and high-resolution images. While the linear-complexity Mamba architecture offers a promising alternative, existing RS applications of Mamba remain limited to supervised tasks on small, domain-specific datasets. To address these challenges, we propose RoMA, a framework that enables scalable self-supervised pretraining of Mamba-based RS foundation models using large-scale, diverse, unlabeled data. RoMA enhances scalability for high-resolution images through a tailored auto-regressive learning strategy, incorporating two key innovations: 1) a rotation-aware pretraining mechanism combining adaptive cropping with angular embeddings to handle sparsely distributed objects with arbitrary orientations, and 2) multi-scale token prediction objectives that address the extreme variations in object scales inherent to RS imagery. Systematic empirical studies validate that Mamba adheres to RS data and parameter scaling laws, with performance scaling reliably as model and data size increase. Furthermore, experiments across scene classification, object detection, and semantic segmentation tasks demonstrate that RoMA-pretrained Mamba models consistently outperform ViT-based counterparts in both accuracy and computational efficiency. The source code and pretrained models will be released at https://github.com/MiliLab/RoMA.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。