AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

CoverPrune: Coverage-Driven Token Pruning for 3D VLMs via Optimal Transport

Peng Ling, Yingda Yin, Lingting Zhu, Weikai Chen, Shengju Qian, Zeyu HU, Xin Wang, Wenming Yang

ECCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Attention-based visual token pruning · 0 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 feature cost term to FST transport cost (ablation)CoverPrune (GS-Reasoner) · VSI-Bench R=20%Overall accuracy: 56.18 -> 59.76Improved
add geometry cost term (ablation)CoverPrune (GS-Reasoner) · VSI-Bench R=20%Overall accuracy: 59.02 -> 59.76Improved
add temporal cost term (ablation)CoverPrune (GS-Reasoner) · VSI-Bench R=20%Overall accuracy: 58.96 -> 59.76Improved
add FST capacity weighting (ablation)CoverPrune (GS-Reasoner) · VSI-Bench R=20%Overall accuracy: 59.59 -> 59.76Improved
use CoverPrune instead of DTC at 10% retentionGS-Reasoner · VSI-Bench R=10%Avg accuracy: 51.66 -> 56.83Improved
use CoverPrune instead of VisionZip at 5% retentionGS-Reasoner · VSI-Bench R=5%Avg accuracy: 46.10 -> 52.66Improved
use CoverPrune instead of DTC at 10% retentionVLM-3R · VSI-Bench R=10%Avg accuracy: 52.64 -> 54.69Improved
use CoverPrune instead of DTC at 5% retentionVLM-3R · VSI-Bench R=5%Avg accuracy: 49.83 -> 51.28Improved
use CoverPrune at 20% retention vs full tokensGS-Reasoner · VSI-BenchAvg accuracy: 64.70 -> 59.76Hurt
use CoverPrune instead of FastVID at 10% retentionGS-Reasoner · Scan2Cap R=10%CIDEr: 85.68 -> 93.84Improved
use CoverPrune-Lite instead of CoverPruneGS-Reasoner · VSI-Bench R=10%Avg accuracy: 56.83 -> 56.94No clear effect
use CoverPrune-Lite instead of CoverPrune for pruning timeGS-Reasoner · VSI-Bench efficiencyPruning time (s): 2.53 -> 0.41Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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