Datasets:

lost-on-campus-project
/

lost-on-campus

Lost on Campus benchmark aims to evaluate ESR of Vision-Language Model (VLM) in large-scale real-world outdoor 3D environments reconstructed by 3D Gaussian Splatting. Our benchmark introduces a unified reasoning-action evaluation framework that seamlessly integrates diagnostic QA for isolated reasoning with closed-loop interactive navigation for active reasoning under multimodal instructions.

Dataset Structure

├── 24_Clark/ 
│ ├── data/ 
│ │ ├── images/
│ │ │ ├── xxx.jpg
│ │ │ ├── ... 
│ │ ├── sparse/ 
│ │ │ ├──0/
│ │ │ │ ├── cameras.bin
│ │ │ │ ├── images.bin
│ │ │ │ ├── points3D.bin
│ │ │ │ ├── points3D.ply
│ ├── outputs/ 
│ │ ├── point_cloud/
│ │ │ ├──iteration_40000/
│ │ │ │ ├── points_cloud.ply
│ │ │ │ ├── ... 
│ │ │ ├── camera.json
│ │ │ ├── cfg.args
│ │ │ ├── input.ply
├── 28_Garland/  
├── .../ |
├── .../ |
├── QA/
│ ├── xxx.json

Downloads last month: 184