Research map / Vision-language and multimodal / Human and embodied modeling
RGBT-GroundBench: Visual Grounding Beyond RGB in Complex Real-World Scenarios
Tianyi Zhao, Jiawen Xi, Linhui Xiao, Junnan Li, Xue Yang, Maoxun Yuan, Xingxing Wei
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 |
|---|---|---|---|
| add AMA (asymmetric modality adaptation) (ablation) | RGBT-VGNet · RefFLIR test | Acc@0.5: 46.19 -> 71.17 | Improved |
| add AMA (ablation) | RGBT-VGNet · RefM3FD test | Acc@0.5: 57.57 -> 72.53 | Improved |
| add LAVS on top of AMA (ablation) | RGBT-VGNet · RefFLIR test | Acc@0.5: 71.17 -> 72.65 | Improved |
| add LAVS on top of AMA (ablation) | RGBT-VGNet · RefFLIR val | Acc@0.5: 74.01 -> 73.68 | Hurt |
| add TPF on top of AMA+LAVS (ablation) | RGBT-VGNet · RefMFAD test | Acc@0.5: 66.62 -> 67.58 | Improved |
| add TPF on top of AMA+LAVS (ablation) | RGBT-VGNet · RefFLIR test | Acc@0.5: 72.65 -> 72.43 | Mixed |
| TPF instead of CMX fusion (ablation) | RGBT-VGNet · RefMFAD test | Acc@0.5: 53.77 -> 67.58 | Improved |
| TPF instead of naive add (ablation) | RGBT-VGNet · RefFLIR test | Acc@0.5: 69.39 -> 72.43 | Improved |
| TPF instead of LIF (ablation) | RGBT-VGNet · RefM3FD test | Acc@0.5: 72.96 -> 74.89 | Improved |
| asymmetric LoRA ranks (16,48) instead of symmetric (32,32) (ablation) | AMA · RefM3FD test | Acc@0.5: 65.20 -> 74.89 | Improved |
| (16,48) instead of more extreme (8,56) (ablation) | AMA · RefM3FD test | Acc@0.5: 66.10 -> 74.89 | Improved |
| add thermal (TIR) modality to RGB input | HiVG-B · RefFLIR test | Acc@0.5: 66.65 -> 69.20 | Improved |
Closest papers by meaning
- When Visual Grounding Meets Gigapixel-level Large-scale Scenes: Benchmark and Approach (CVPR 2024 · similarity 0.89)
- SeeGround: See and Ground for Zero-Shot Open-Vocabulary 3D Visual Grounding (CVPR 2025 · similarity 0.89)
- GroundVLP: Harnessing Zero-Shot Visual Grounding from Vision-Language Pre-training and Open-Vocabulary Object Detection (AAAI 2024 · similarity 0.88)
- From Objects to Anywhere: A Holistic Benchmark for Multi-level Visual Grounding in 3D Scenes (NeurIPS 2025 · similarity 0.88)
- SimVG: A Simple Framework for Visual Grounding with Decoupled Multi-modal Fusion (NeurIPS 2024 · similarity 0.87)
- 3EED: Ground Everything Everywhere in 3D (NeurIPS 2025 · similarity 0.87)
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.
