AI Paper Atlas

Research map / Point clouds / Segmentation and detection

MR-COSMO: Visual-Text Memory Recall and Direct CrOSs-MOdal Alignment Method for Query-Driven 3D Segmentation

Chade Li, Pengju Zhang, Yihong Wu

AAAI 2026 · Point clouds › Segmentation and detection · approach: Point cloud semantic segmentation · 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 Memory Module for Feature Pairs (ablation)MR-COSMO · Instruct3DAcc: 31.9 -> 33.8Improved
add Memory Module for Feature Pairs (ablation)MR-COSMO · Instruct3DmIoU: 27.4 -> 28.5Improved
add Direct Cross-Modal Alignment to baseline (ablation)MR-COSMO · Instruct3DmIoU: 26.4 -> 27.4Improved
add Memory Module to baseline alone (ablation)MR-COSMO · Instruct3DmIoU: 26.4 -> 27.5Improved
combine DCMA and Memory Module (ablation)MR-COSMO · Instruct3DmIoU: 26.4 -> 28.5Improved
add MLP point features to voxel encoder (ablation)MR-COSMO · Instruct3DmIoU: 27.7 -> 28.5Improved
use image-voxel and text-point matching instead of reversed (ablation)MR-COSMO · Instruct3DmIoU: 27.9 -> 28.5Improved
add Alignment Constrains Block (contrastive) (ablation)MR-COSMO · Instruct3DmIoU: 28.0 -> 28.5Improved
add loss on bounding boxes (ablation)MR-COSMO · Instruct3DmIoU: 28.4 -> 28.5No clear effect
replace Mamba with Transformer (ablation)MR-COSMO · Instruct3DmIoU: 28.5 -> 28.4Hurt
replace ResNet50 with ViT (ablation)MR-COSMO · Instruct3DmIoU: 28.5 -> 28.7Mixed
scale LLaMA2 from 7B to 13B (ablation)MR-COSMO · Instruct3DmIoU: 28.5 -> 28.6No clear effect

Later papers that compared with this method

1 later accepted paper reports MR-COSMO as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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