Research map / Vision-language and multimodal / Human and embodied modeling
Learning Cross-Modal Context Graph for Visual Grounding
Yongfei Liu, Bo Wan, Xiaodan Zhu, Xuming He
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use the cross-modal graph approach with ResNet-101 instead of DDPN | visual grounding · Flickr30K Entities test | accuracy (%): 73.30 -> 76.74 | Improved |
| use the cross-modal graph approach with ResNet-50 instead of SeqGROUND | visual grounding · Flickr30K Entities test | accuracy (%): 61.60 -> 67.90 | Improved |
| add phrase graph network (ablation) | visual grounding baseline · Flickr30K Entities val | accuracy (%): 73.46 -> 74.40 | Improved |
| add proposal pruning guided by PGN (ablation) | baseline + PGN · Flickr30K Entities val | accuracy (%): 74.40 -> 75.50 | Improved |
| add visual object graph network (ablation) | baseline + PGN + proposal pruning · Flickr30K Entities val | accuracy (%): 75.50 -> 75.85 | Improved |
| add structured prediction with global relation matching (ablation) | baseline + PGN + proposal pruning + VOGN · Flickr30K Entities val | accuracy (%): 75.85% to 76.19% | Improved |
| remove phrase relation features (ablation) | phrase graph network · Flickr30K Entities val | accuracy (%): 74.40 -> 74.11 | Hurt |
| remove phrase relation features (ablation) | PGN + proposal pruning · Flickr30K Entities val | accuracy (%): 75.50 -> 75.32 | Hurt |
| remove phrase and visual relation features (ablation) | PGN + proposal pruning + VOGN · Flickr30K Entities val | accuracy (%): 75.85 -> 75.44 | Hurt |
| reduce proposals per phrase from K=10 to K=5 (ablation) | cross-modal graph model · Flickr30K Entities val | accuracy (%): 76.19% to 74.97% | Hurt |
| increase proposals per phrase from K=10 to K=20 (ablation) | cross-modal graph model · Flickr30K Entities val | accuracy (%): 76.19 -> 76.07 | No clear effect |
Closest papers by meaning
- Look Around Before Locating: Considering Content and Structure Information for Visual Grounding (AAAI 2025 · similarity 0.90)
- SimVG: A Simple Framework for Visual Grounding with Decoupled Multi-modal Fusion (NeurIPS 2024 · similarity 0.90)
- Connecting the Dots: Training-Free Visual Grounding via Agentic Reasoning (AAAI 2026 · similarity 0.90)
- GroundingME: Exposing the Visual Grounding Gap in MLLMs through Multi-Dimensional Evaluation (CVPR 2026 · similarity 0.90)
- GeM-VG: Towards Generalized Multi-image Visual Grounding with Multimodal Large Language Models (AAAI 2026 · similarity 0.90)
- Cross-Modal Match for Language Conditioned 3D Object Grounding (AAAI 2024 · similarity 0.90)
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.
