针对超高清遥感图像，实现精准且低算力的视觉令牌压缩。
UHR-BAT: Budget-Aware Token Compression Vision-Language model for Ultra-High-Resolution Remote Sensing
- 基于文本引导的多尺度重要性评估，动态筛选关键视觉令牌。
- 在多个基准上超越现有方法，保持高精度同时降低计算开销。
- 适合处理大范围遥感图像中的小目标检测与分析任务。
超高清遥感图像同时包含千米级上下文和仅占几个像素的关键证据，其巨大空间尺度导致视觉令牌数量呈二次增长，阻碍小目标信息提取。以往方法采用直接下采样、密集切片或全局 top-k 精简，要么牺牲查询关键细节，要么带来不可控的计算开销。本文提出 UHR-BAT，一种查询引导、区域忠实的令牌压缩框架，在严格上下文预算下高效选择视觉令牌。具体而言，利用文本引导的多尺度重要性估计，有效解决精确且低成本特征提取难题；通过引入区域保留与合并策略，减少视觉令牌冗余，进一步降低计算成本。实验表明，UHR-BAT 在多个基准上达到当前最优性能。代码将发布于 https://github.com/Yunkaidang/UHR。
原文摘要 · Abstract (English)
Ultra-high-resolution (UHR) remote sensing imagery couples kilometer-scale context with query-critical evidence that may occupy only a few pixels. Such vast spatial scale leads to a quadratic explosion of visual tokens and hinders the extraction of information from small objects. Previous works utilize direct downsampling, dense tiling, or global top-k pruning, which either compromise query-critical image details or incur unpredictable compute. In this paper, we propose UHR-BAT, a query-guided and region-faithful token compression framework to efficiently select visual tokens under a strict context budget. Specifically, we leverage text-guided, multi-scale importance estimation for visual tokens, effectively tackling the challenge of achieving precise yet low-cost feature extraction. Furthermore, by introducing region-wise preserve and merge strategies, we mitigate visual token redundancy, further driving down the computational budget. Experimental results show that UHR-BAT achieves state-of-the-art performance across various benchmarks. Code will be available at https://github.com/Yunkaidang/UHR.
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。