Search papers, labs, and topics across Lattice.
This paper introduces DualComp, a training-free visual token reduction framework tailored for ultra-high-resolution remote sensing imagery, which addresses the computational bottlenecks of applying MLLMs to this domain. DualComp uses a dual-stream architecture with a semantic stream that aggressively prunes background tokens and a geometric stream that preserves spatial topology. Experiments on XLRS-Bench show that DualComp achieves both improved efficiency and accuracy in remote sensing interpretation tasks.
Achieve faster and more accurate remote sensing interpretation by intelligently pruning visual tokens based on task-specific semantic and geometric importance, without any training.
Multimodal Large Language Models (MLLMs) have demonstrated immense potential in Earth observation. However, the massive visual tokens generated when processing Ultra-High-Resolution (UHR) imagery introduce prohibitive computational overhead, severely bottlenecking their inference efficiency. Existing visual token compression methods predominantly adopt static and uniform compression strategies, neglecting the inherent "Semantic-Geometric Duality" in remote sensing interpretation tasks. Specifically, object semantic tasks focus on the abstract semantics of objects and benefit from aggressive background pruning, whereas scene geometric tasks critically rely on the integrity of spatial topology. To address this challenge, we propose DualComp, a task-adaptive dual-stream token compression framework. Dynamically guided by a lightweight pre-trained router, DualComp decouples feature processing into two dedicated pathways. In the object semantic stream, the Spatially-Contiguous Semantic Aggregator (SCSA) utilizes size-adaptive clustering to aggregates redundant background while protecting small object. In the scene geometric stream, the Instruction-Guided Structure Recoverer (IGSR) introduces a greedy path-tracing topology completion mechanism to reconstruct spatial skeletons. Experiments on the UHR remote sensing benchmark XLRS-Bench demonstrate that DualComp accomplishes high-fidelity remote sensing interpretation at an exceptionally low computational cost, achieving simultaneous improvements in both efficiency and accuracy.
