Table 1 Review of the main elements of the most popular RGB-D dataset.
From: UASOL, a large-scale high-resolution outdoor stereo dataset
Dataset | Scale | Resolution | Modality | Outdoor | Video | |||
|---|---|---|---|---|---|---|---|---|
Frames | Scenes | RGB | Depth | Point Cloud | ||||
SCENENET RGB-D1 | 5 M | 57 | 320 × 240 | ✓ | ✓ | ✗ | ✗ | ✓ |
MATTERPORT 3D24 | 200 k | 90 | 1280 × 1024 | ✓ | ✓ | ✓ | ✗ | ✓ |
STANFORD 2D-3D25 | 70 k | 270 | 1080 × 1080 | ✓ | ✓ | ✓ | ✗ | ✓ |
SCENE NN26 | 2 k–10 k each scene | 100 | 640 × 480 | ✓ | ✓ | ✗ | ✗ | ✓ |
MICROSOFT RGB-D27 | 500–1 k each seq. | 7 | 640 × 480 | ✓ | ✓ | ✗ | ✗ | ✓ |
VIDRILO28 | 22 k | 10 | 640 × 480 | ✓ | ✗ | ✓ | ✗ | ✓ |
SUN RGBD29 | 10 k | 270 | 1080 × 1080 | ✓ | ✓ | ✓ | ✗ | ✗ |
NYU-D V230 | 1.5 k | 464 | 640 × 480 | ✓ | ✓ | ✗ | ✗ | ✓ |
B3DO31 | 849 | 75 | 1080 × 1080 | ✓ | ✓ | ✗ | ✗ | ✗ |
SYNTHIA2 | 200 k | 960 × 720 | ✓ | ✓ | ✗ | ✓ | ✗ | |
KITTI5 | 1.6 k | 400 | 1392 × 512 | ✓ | ✓ | ✓ | ✓ | ✗ |
ETH3D4 | 898 | 25 | ✓ | ✓ | ✗ | ✓ | ✓ | |
MAKE 3D3 | 534 | 1280 × 1024 | ✓ | ✓ | ✗ | ✓ | ✓ | |
Tanks and Temples6 | 147791 | 14 | 1600 × 1200 | ✓ | ✓ | ✓ | ✗ | ✓ |
Middlebury Dataset7 | 40 k | 33 | 2964 × 1988 | ✓ | ✓ | ✗ | ✗ | ✗ |
OURS | 482736 | 33 | 2208 × 1242 | ✓ | ✓ | ✗ | ✓ | ✓ |
