Skip to content

Latest commit

 

History

History
52 lines (42 loc) · 1.75 KB

File metadata and controls

52 lines (42 loc) · 1.75 KB

Data Processing

🔽 Download 3D Scenes and Annotations

To generate bird’s-eye-view (BEV) images and project 3D object marks, you need to download three indoor scene datasets:

After downloading, organize the datasets in the following directory structure:

.
├── arkitscenes
│   ├── data
│   │   ├── ...
│   │   └── {scene_id}_3dod_mesh_wo_ceiling.ply          # point cloud files
│   ├── gt_annotation                                     # ground truth annotations
│   │   └── {scene_id}_3dod_annotation.json
│   └── detection_annotation                              # noisy detection annotations
│
├── scannet
│   ├── data
│   │   ├── ...
│   │   └── {scene_id}_vh_clean_2.ply
│   ├── gt_annotation
│   │   ├── {scene_id}.aggregation.json
│   │   └── {scene_id}_vh_clean_2.0.010000.segs.json
│   └── detection_annotation
│
├── scannetpp
│   ├── data
│   │   ├── ...
│   │   └── {scene_id}_mesh_aligned_0.05.ply
│   ├── gt_annotation
│   │   ├── {scene_id}_segments_anno.json
│   │   └── {scene_id}_segments.json
│   └── detection_annotation
└── ...

Generate BEV Images for VSI-Bench Evaluation

Use the script below to generate BEV images and save 3D object bounding boxes as JSON files:

python generate_bev_images.py --dataset all

We are currently streamlining the data generation pipeline for training and evaluation, as the original datasets are too large to be released directly.