Research map / 3D vision / Human pose estimation
GeoX-Bench: Benchmarking Cross-View Geo-Localization and Pose Estimation Capabilities of Large Multimodal Models
Yushuo Zheng, Jiangyong Ying, Huiyu Duan, Chunyi Li, Zicheng Zhang, Jing Liu, Xiaohong Liu, Guangtao Zhai
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 |
|---|---|---|---|
| LoRA instruction tuning on GeoX-Bench | InternVL3-8B · Pose estimation (fixed) | accuracy: 25.36 -> 55.36 | Improved |
| LoRA instruction tuning on GeoX-Bench | InternVL3-8B · Average (7 tasks) | accuracy: 40.08 -> 71.35 | Improved |
| LoRA instruction tuning on GeoX-Bench | InternVL2-8B · Pose estimation (fixed) | accuracy: 26.27 -> 25.00 | No clear effect |
| LoRA instruction tuning on GeoX-Bench | Qwen2.5VL-7B · Average (7 tasks) | accuracy: 38.56 -> 67.38 | Improved |
| LoRA instruction tuning on GeoX-Bench | Qwen2VL-7B · Average (7 tasks) | accuracy: 36.03 -> 68.96 | Improved |
| scale model up (2B to 72B) | LMMs · Option bias | normalized entropy: 0.4412 (Qwen2VL-2B) -> 0.8216 (Qwen2VL-72B) | Improved |
| scale beyond about 32B parameters | pretrained LMMs · Average (7 tasks) | accuracy: 43.76 (Qwen2.5VL-32B) -> 48.30 (72B) | No clear effect |
| scale model up on pose estimation | pretrained LMMs · Pose estimation | accuracy: no open model above 30% | No clear effect |
Closest papers by meaning
- MMLandmarks: a Cross-View Instance-Level Benchmark for Geo-Spatial Understanding (CVPR 2026 · similarity 0.89)
- GeoBayes: Probabilistic Image Geo-Localization Inference via Sequential Bayesian Updating (AAAI 2026 · similarity 0.89)
- ERGeoBench: A Comprehensive Benchmark for Embodied Reasoning and Geo-localization in Multimodal Large Language Models (ICML 2026 · similarity 0.88)
- UrBench: A Comprehensive Benchmark for Evaluating Large Multimodal Models in Multi-View Urban Scenarios (AAAI 2025 · similarity 0.88)
- Aligning Geometric Spatial Layout in Cross-View Geo-Localization via Feature Recombination (AAAI 2024 · similarity 0.87)
- GeoPixel: Pixel Grounding Large Multimodal Model in Remote Sensing (ICML 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.
