Research map / Vision-language and multimodal / Multimodal LLMs
CVSBench: A Comprehensive Benchmark for Cross-view Spatial Reasoning and Dreaming
ruixun liu, Lingyu Zhang, Lanxuan Xue, Kaiyu Li, Bowen Fu, Xiangyong Cao
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 inference-time CoT prompting (ablation) | Qwen3-VL-4B · CVUSA-subset grounding | mIoU (%): 3.68 -> 6.01 | Improved |
| add inference-time CoT prompting (ablation) | Qwen3-VL-4B · FOV-subset G2S | accuracy (%): 26.70 -> 35.60 | Improved |
| add inference-time CoT prompting (ablation) | Qwen3-VL-4B · FOV-subset View-Arrow | accuracy (%): 40.80 -> 21.90 | Hurt |
| add inference-time CoT prompting (ablation) | Qwen3-VL-4B · FOV-subset View-Image | accuracy (%): 27.10 -> 23.60 | Hurt |
| add inference-time CoT prompting (ablation) | Qwen3-VL-4B · CVUSA-subset S2G | accuracy (%): 62.40 -> 61.70 | Hurt |
| add Structured Scene CoT with SFT (ablation) | Qwen3-VL-4B · FOV-subset G2S facade | accuracy (%): 18.4 -> 41.3 | Improved |
| add Spatial Imagination CoT with SFT (ablation) | Qwen3-VL-4B · CVUSA-subset G2S distance | accuracy (%): 84.0 -> 89.1 | Improved |
| add Spatial Imagination CoT with SFT (ablation) | Qwen3-VL-4B · CVUSA-subset G2S roof | accuracy (%): 63.2 -> 46.5 | Hurt |
| add RL after Structured Scene SFT (ablation) | Qwen3-VL-4B · FOV-subset G2S facade | accuracy (%): 41.3 -> 25.5 | Hurt |
| add RL after Structured Scene SFT (ablation) | Qwen3-VL-4B · CVUSA-subset G2S connection | accuracy (%): 28.1 -> 37.2 | Improved |
| add 3D miniature auxiliary view (ablation) | Qwen3-VL-4B · FOV-subset S2G color | accuracy (%): 24.5 -> 36.9 | Improved |
| add 3D miniature auxiliary view (ablation) | Qwen3-VL-4B · FOV-subset G2S roof | accuracy (%): 9.9 -> 18.7 | Improved |
Closest papers by meaning
- SpatiaLab: Can Vision–Language Models Perform Spatial Reasoning in the Wild? (ICLR 2026 · similarity 0.92)
- SpinBench: Perspective and Rotation as a Lens on Spatial Reasoning in VLMs (ICLR 2026 · similarity 0.92)
- Spatial-DISE: A Unified Benchmark for Evaluating Spatial Reasoning in Vision-Language Models (ICLR 2026 · similarity 0.91)
- InternSpatial: A Comprehensive Dataset for Spatial Reasoning in Vision-Language Models (ICLR 2026 · similarity 0.91)
- MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence (ICLR 2026 · similarity 0.91)
- SpatialRGPT: Grounded Spatial Reasoning in Vision-Language Models (NeurIPS 2024 · similarity 0.91)
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.
