AI Paper Atlas

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

Vision-Language Interactive Relation Mining for Open-Vocabulary Scene Graph Generation

Yukuan Min, Muli Yang, Jinhao Zhang, Yuxuan Wang, Aming Wu, Cheng Deng

ICCV 2025 · Vision-language and multimodal › Human and embodied modeling · approach: Scene graph generation · 4 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 learnable context tokens (ablation)VL-IRM · VG150 Novel+baseR@100: 16.67 -> 18.48Improved
add learnable context tokens (vs frozen) (ablation)VL-IRM · VG150 Novel+baseR@100: 15.36 -> 18.48Improved
add dynamic fitting debiasing (ablation)VL-IRM · VG150 Novel+baseR@100: 16.24 -> 18.48Improved
add dynamic fitting debiasing (ablation)VL-IRM · VG150 NovelR@100: 10.23 -> 13.45Improved
use dynamic fitting instead of static LA (ablation)VL-IRM · VG150 Novel+baseR@100: 12.58 -> 18.48Improved
use dynamic fitting instead of PGLA (ablation)VL-IRM · VG150 Novel+baseR@100: 13.42 -> 18.48Improved
add background and foreground pseudo labels vs foreground only (ablation)VL-IRM · VG150 Novel+BaseR@100: 19.12 -> 19.71Improved
add pseudo labeling (Bg & fg) (ablation)VL-IRM · VG150 NovelR@100: 13.45 -> 14.32Improved
add relation language generator (lambda>0) (ablation)VL-IRM · VG150 NovelperformanceImproved
combine VL-IRM with OVSGTROVSGTR · VG SGDet extra data, Novel+baseR@100: 23.9 -> 25.0Improved
combine VL-IRM with OVSGTR-TOVSGTR-T · VG150 OvD+R-SGG Novel relationR@100: 11.19 -> 15.26Improved
VL-IRM vs OVSGTR on OvD+R-SGG novel R@50OVSGTR-T · VG150 OvD+R-SGG NovelR@50: 9.20 -> 12.27Improved

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