AI Paper Atlas

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

Visual Grounding for Object Questions

Martin Nicolas Everaert, Xiruo Liu, Hiroyuki Takeda, Raja Bala, Vivek Yadav, Vidya Narayanan

CVPR 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Visual grounding transformers

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
rephrase visual question as general object questionSoTA grounding models (UnifiedIO-XL) · VizWiz images, specific visual evidence subsetgIoU (%): 52.2 -> 37.2Hurt
train lightweight CLIPSeg-style model on VGOQ data jointly with RES and VQA grounding dataOur LW (1.77M parameters) · ABO-VGOQ val specific visual evidencegIoU (%): 12.4 (UnifiedIO-XL) -> 39.5Improved
compare 1.77M lightweight model with Qwen3-VL-8BOur LW · ABO-VGOQ test specific visual evidencegIoU (%): 30.3 -> 32.5 (Qwen3-VL-4B 32.7)No clear effect
compare lightweight model with Qwen3-VL-8B on VizWiz-VGOQ related-to-questionOur LW · VizWiz-VGOQ val-RTQgIoU (%): 49.4 -> 64.1Improved
compare lightweight model on Toloka VQA grounding (not trained on)Our LW · Toloka test-publicgIoU (%): 71.4 (Qwen3-VL-8B) -> 12.6Hurt
lightweight model versus UnifiedIO-XL on VizWiz-VQA-Grounding VQOur LW · VizWiz-VQA-Grounding valgIoU (%): 68.1 -> 67.2Hurt
finetune only on specific-visual-evidence samples (ablation)Our LW · ABO-VGOQ / VizWiz-VGOQ specific evidencegIoU (percentage points): +1.7 to +7.4Improved
scale UnifiedIO from Small to XLUnifiedIO · ABO-VGOQ val specific visual evidencegIoU (%): 12.3 -> 12.4No clear effect

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