AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

When Large Vision-Language Model Meets Large Remote Sensing Imagery: Coarse-to-Fine Text-Guided Token Pruning

Junwei Luo, Yingying Zhang, Xue Yang, Kang Wu, Qi Zhu, Lei Liang, Jingdong Chen, Yansheng Li

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Attention-based visual token pruning · 32 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add RFM text-guided pruning with DIP to LLaVA-1.5LLaVA-1.5 · Avg of 4 datasetsAvg accuracy: 27.30 -> 31.59Improved
add RFM text-guided pruning with DIP to LLaVA-NextLLaVA-Next-Qwen2 · Avg of 4 datasetsAvg accuracy: 30.80 -> 32.16Improved
prune 75% of vision tokens by RFM without DIPLLaVA-Next-p25 · MME-RealWorld-RSaccuracy: 39.65 -> 41.28Improved
prune 90% instead of 75% of vision tokens (ablation)w/ prune (Ours) · MME-RealWorld-RSaccuracy: 41.28 -> 39.41Hurt
use RFM guidance instead of CLIP similarity for pruning (ablation)anyres-p25 · MME-RealWorld-RSaccuracy: 36.86 -> 41.28Improved
use RFM guidance instead of RemoteCLIP similarity (ablation)anyres-p25 · MME-RealWorld-RSaccuracy: 38.77 -> 41.28Improved
add PruMerge++ pruning (ablation)anyres-p25 · MME-RealWorld-RSaccuracy: 39.65 -> 34.86Hurt
add MSE loss on high-res tokens to KL distillation (ablation)RFM · MME-RealWorld-RSaccuracy: 41.20 -> 41.89Improved
add KL loss to MSE-only distillation (ablation)RFM · MME-RealWorld-RSaccuracy: 39.73 -> 41.89Improved
use dynamic DIP depth instead of fixed 5 layers (ablation)DIP inference · MME-RealWorld-RSaccuracy: 39.91 -> 41.89Improved
use select instead of concat of tokens from traversed DIP layers (ablation)RFM pruning · MME-RealWorld-RSaccuracy: 41.39 -> 41.89Improved
apply token pruning during SFT training (ablation)RFM training · MME-RealWorld-RSaccuracy: 41.89 -> 39.09Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy