AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

VIRTUE: Visual-Interactive Text-Image Universal Embedder

Wei-Yao Wang, Kazuya Tateishi, Qiyu Wu, Shusuke Takahashi, Yuki Mitsufuji

ICLR 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Scene graph generation · 1 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 SAM-2 segmentation embeddings (VIRTUE) to a VLM embedder with same dataQwen2-VL-2B VLM2Vec · MMEB overallprecision@1: 59.7 -> 64.8Improved
add segmentation embeddings (VIRTUE)7B embedders · MMEB overallprecision@1: 66.6 -> 68.6Improved
add segmentation embeddings (VIRTUE)VLM2Vec-2B · SCaR overallprecision@1: 24.1 -> 30.4Improved
add segmentation embeddings (VIRTUE)7B embedders · SCaR overallprecision@1: 26.3 -> 27.8Improved
fine-tune on SCaR-trainVIRTUE-2B · SCaR overallprecision@1: 30.4 -> 56.2Improved
crop the region of interest instead of using visual prompt (ablation)VIRTUE-2B and other models · SCaR overallprecision@1: 30.4 -> 24.8Hurt
add red circle visual hint to CLIP-based ReCLIPReCLIP · SCaR overallprecision@1: 18.1 -> 19.0No clear effect
remove task instruction from query (ablation)VIRTUE-2B · MMEB overallprecision@1: 59.2 -> 64.8Improved
use SAM-2 instead of Prompter or cropping (ablation)VIRTUE-2B · SCaR overallprecision@1: 22.7 (Prompter), 25.9 (Cropped) -> 30.4Improved
increase SAM2.1 size from B+ to L (ablation)VIRTUE-2B · SCaR overallprecision@1: 30.4 -> 27.9Hurt
increase segmentation embedding length from 256 to 1024 (ablation)VIRTUE-2B · MMEB vs SCaRprecision@1: MMEB 64.8 -> 60.2; SCaR 30.4 -> 35.1Mixed
unfreeze SAM-2 during fine-tuning (ablation)VIRTUE-2B · SCaR overallprecision@1: 30.4 -> 25.8Hurt

Later papers that compared with this method

1 later accepted paper reports VIRTUE-2B as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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