Dataset Viewer
Auto-converted to Parquet Duplicate
image
imagewidth (px)
640
640
depth
imagewidth (px)
640
640
raw_depth
imagewidth (px)
640
640
label894
imagewidth (px)
640
640
label40
imagewidth (px)
640
640
label13
imagewidth (px)
640
640
instance
imagewidth (px)
640
640
index
int32
3
1.44k
scene
stringlengths
9
20
scene_type
stringclasses
27 values
raw_rgb_filename
stringlengths
43
56
3
office_0003
office
office_0003/r-1294851609.500330-2975113445.ppm
4
office_0003
office
office_0003/r-1294851614.847235-3293159158.ppm
5
office_0004
office
office_0004/r-1294851494.927741-395092604.ppm
6
office_0004
office
office_0004/r-1294851498.018189-581119341.ppm
7
office_0004
office
office_0004/r-1294851501.026030-761145216.ppm
8
office_0004
office
office_0004/r-1294851516.824994-1709281491.ppm
10
office_0006
office
office_0006/r-1294851326.690952-3186543767.ppm
11
office_0006
office
office_0006/r-1294851340.299448-4002661067.ppm
12
office_0006
office
office_0006/r-1294851348.033010-171760471.ppm
13
office_0006
office
office_0006/r-1294851359.399276-853858508.ppm
19
office_0009
office
office_0009/r-1294851157.595315-1631020159.ppm
20
office_0009
office
office_0009/r-1294851162.927053-1951066159.ppm
22
office_0011
office
office_0011/r-1294850923.235047-453900922.ppm
23
office_0011
office
office_0011/r-1294850942.497012-1610067097.ppm
24
office_0011
office
office_0011/r-1294850951.737763-2164146734.ppm
25
office_0011
office
office_0011/r-1294850959.741116-2644215734.ppm
26
office_0012
office
office_0012/r-1294850888.528910-2666568930.ppm
27
office_0012
office
office_0012/r-1294850910.101702-3960754943.ppm
44
bathroom_0002
bathroom
bathroom_0002/r-1295148733.248304-3836198014.ppm
45
bathroom_0002
bathroom
bathroom_0002/r-1295148736.261033-4016223889.ppm
48
bathroom_0005
bathroom
bathroom_0005/r-1295837594.698575-3478376057.ppm
49
bathroom_0006
bathroom
bathroom_0006/r-1295839279.324567-1475688615.ppm
50
living_room_0000
living_room
bedroom_0010/r-1295834875.424236-3509809043.ppm
51
living_room_0000
living_room
bedroom_0010/r-1295834884.556643-4057887818.ppm
52
living_room_0000
living_room
bedroom_0010/r-1295834897.568688-543032647.ppm
53
living_room_0000
living_room
bedroom_0010/r-1295834899.399739-653048459.ppm
54
living_room_0000
living_room
bedroom_0010/r-1295834904.681547-965093309.ppm
55
living_room_0000
living_room
bedroom_0010/r-1295834909.934990-1285139309.ppm
58
bedroom_0012
bedroom
bedroom_0012/r-1295837389.763849-4067510682.ppm
64
bedroom_0014
bedroom
bedroom_0014/r-1295838120.106836-643168689.ppm
65
bedroom_0014
bedroom
bedroom_0014/r-1295838125.816144-985217851.ppm
66
bedroom_0014
bedroom
bedroom_0014/r-1295838135.139621-1545298351.ppm
67
bedroom_0015
bedroom
bedroom_0015/r-1295838185.079525-246761730.ppm
68
bedroom_0016
bedroom
bedroom_0016/r-1295838249.636228-4119318330.ppm
69
bedroom_0016
bedroom
bedroom_0016/r-1295838254.007359-64385534.ppm
70
bedroom_0016
bedroom
bedroom_0016/r-1295838268.133252-934510597.ppm
71
bedroom_0016
bedroom
bedroom_0016/r-1295838278.600399-1562600872.ppm
72
bedroom_0016
bedroom
bedroom_0016/r-1295838290.694714-2288705234.ppm
73
bedroom_0017
bedroom
bedroom_0017/r-1295838401.589571-351726867.ppm
74
bedroom_0017
bedroom
bedroom_0017/r-1295838411.753437-961814555.ppm
75
bedroom_0017
bedroom
bedroom_0017/r-1295838427.150000-1885947380.ppm
80
bedroom_0019
bedroom
bedroom_0019/r-1295839353.583350-1635361582.ppm
81
bedroom_0019
bedroom
bedroom_0019/r-1295839360.302178-2039419657.ppm
82
bedroom_0019
bedroom
bedroom_0019/r-1295839369.497065-2591499007.ppm
83
bedroom_0019
bedroom
bedroom_0019/r-1295839387.999466-3701658569.ppm
92
bookstore_0002
bookstore
bookstore_0001d/r-1295525983.211876-3836874081.ppm
93
bookstore_0002
bookstore
bookstore_0001d/r-1295526013.918171-1384171573.ppm
94
bookstore_0002
bookstore
bookstore_0001d/r-1295526036.795136-2756368798.ppm
95
bookstore_0002
bookstore
bookstore_0001d/r-1295526061.964206-4266585860.ppm
96
bookstore_0002
bookstore
bookstore_0001d/r-1295526070.197628-465689577.ppm
97
bookstore_0002
bookstore
bookstore_0001e/r-1295526090.493155-1677863802.ppm
98
bookstore_0002
bookstore
bookstore_0001e/r-1295526102.635200-2411969314.ppm
99
bookstore_0002
bookstore
bookstore_0001e/r-1295526113.375695-3056061889.ppm
100
bookstore_0002
bookstore
bookstore_0001e/r-1295526118.438466-3360105589.ppm
101
bookstore_0002
bookstore
bookstore_0001e/r-1295526126.300145-3832173439.ppm
102
bookstore_0002
bookstore
bookstore_0001e/r-1295526136.498732-149294118.ppm
103
bookstore_0002
bookstore
bookstore_0001e/r-1295526144.037340-601359093.ppm
104
bookstore_0000
bookstore
bookstore_0001f/r-1295526184.351568-3019706681.ppm
105
bookstore_0000
bookstore
bookstore_0001f/r-1295526233.028053-1645159135.ppm
106
bookstore_0000
bookstore
bookstore_0001f/r-1295526257.564665-3117370735.ppm
107
bookstore_0000
bookstore
bookstore_0001f/r-1295526286.470062-556652701.ppm
108
bookstore_0000
bookstore
bookstore_0001g/r-1295526347.740647-4233181126.ppm
109
bookstore_0000
bookstore
bookstore_0001g/r-1295526368.584325-1188393518.ppm
110
bookstore_0000
bookstore
bookstore_0001g/r-1295526379.316706-1832486093.ppm
111
bookstore_0000
bookstore
bookstore_0001g/r-1295526395.720943-2816627543.ppm
112
bookstore_0000
bookstore
bookstore_0001g/r-1295526404.350512-3334702005.ppm
113
bookstore_0000
bookstore
bookstore_0001h/r-1295526472.193420-3110319772.ppm
114
bookstore_0002
bookstore
bookstore_0001i/r-1295526567.432592-235206567.ppm
115
bookstore_0002
bookstore
bookstore_0001j/r-1295526629.452763-3955741317.ppm
116
bookstore_0002
bookstore
bookstore_0001j/r-1295526689.798362-3281294396.ppm
120
cafe_0001
cafe
cafe_0001a/r-1295527104.556524-2397067408.ppm
121
cafe_0001
cafe
cafe_0001b/r-1295527252.453655-2680408166.ppm
122
cafe_0001
cafe
cafe_0001c/r-1295527289.584149-613761145.ppm
123
cafe_0001
cafe
cafe_0001c/r-1295527298.224383-1131835607.ppm
124
cafe_0001
cafe
cafe_0001c/r-1295527310.222003-1851939107.ppm
130
kitchen_0006
kitchen
kitchen_0006/r-1295406791.964010-1059809939.ppm
135
kitchen_0008
kitchen
kitchen_0008/r-1295836629.443998-2913585842.ppm
136
kitchen_0008
kitchen
kitchen_0008/r-1295836636.752978-3351648804.ppm
138
kitchen_0010
kitchen
kitchen_0010/r-1295838059.499554-1301613310.ppm
139
kitchen_0010
kitchen
kitchen_0010/r-1295838064.593117-1607657297.ppm
140
kitchen_0010
kitchen
kitchen_0010/r-1295838072.833872-2101728310.ppm
141
kitchen_0010
kitchen
kitchen_0010/r-1295838081.354717-2593799035.ppm
142
kitchen_0011
kitchen
kitchen_0011a/r-1295838957.743376-3655751870.ppm
143
kitchen_0011
kitchen
kitchen_0011a/r-1295838964.572961-4065810808.ppm
144
kitchen_0011
kitchen
kitchen_0011a/r-1295838984.400468-959014287.ppm
145
kitchen_0011
kitchen
kitchen_0011b/r-1295839016.418242-2881290574.ppm
146
living_room_0004
living_room
living_room_0004/r-1295148523.379492-4129290090.ppm
147
living_room_0004
living_room
living_room_0004/r-1295148531.722968-334394669.ppm
148
living_room_0004
living_room
living_room_0004/r-1295148545.864139-1182516569.ppm
149
living_room_0004
living_room
living_room_0004/r-1295148551.899603-1544568606.ppm
150
living_room_0004
living_room
living_room_0004/r-1295148560.224170-2044640481.ppm
151
living_room_0005
living_room
living_room_0005/r-1295148605.597922-460062748.ppm
152
living_room_0006
living_room
living_room_0006/r-1295148648.277565-3032432473.ppm
156
living_room_0010
living_room
living_room_0010/r-1295836459.625028-1314055909.ppm
157
living_room_0010
living_room
living_room_0010/r-1295836493.296511-3334346284.ppm
158
living_room_0010
living_room
living_room_0010/r-1295836504.802117-4022445184.ppm
159
living_room_0010
living_room
living_room_0010/r-1295836508.264486-4232475371.ppm
160
living_room_0011
living_room
living_room_0011/r-1295837631.797054-1409728748.ppm
161
living_room_0011
living_room
living_room_0011/r-1295837645.944195-2257850648.ppm
162
living_room_0011
living_room
living_room_0011/r-1295837653.368703-2703914761.ppm
End of preview. Expand in Data Studio

NYU Depth v2: Densely Labeled Indoor RGB-D Frames

Task Dataset Format Classes Splits License

Unofficial redistribution of the NYU Depth v2 labeled subset (1,449 frames) as viewer-ready Parquet: RGB, filled and raw depth, the original 894-class labels, the standard 40- and 13-class labels, instance masks, and scene names, with the official 795 / 654 split.

Disclaimer

This repository is not an official release of NYU Depth v2.

NYU Depth v2 was created by Nathan Silberman, Derek Hoiem, Pushmeet Kohli and Rob Fergus (New York University), who retain all copyright and intellectual property rights. This repository does not claim ownership of any images, annotations, or metadata.

No license was specified by the original authors. The project page only asks users to cite the paper below. This copy is an unofficial redistribution for research purposes, with full attribution. If you are an author or rights holder and want this repository removed or changed, please open a discussion on this repository and it will be addressed promptly.

The 40-class mapping (Gupta et al., CVPR 2013) and 13-class mapping (as used by Couprie et al. and SceneNet) are taken from Ankur Handa's nyuv2-meta-data (no license specified).


Dataset Overview

NYU Depth v2 contains video sequences of 464 indoor scenes recorded with a Microsoft Kinect in three US cities. The labeled subset (this repository) has 1,449 RGB-D frames (640x480) with dense per-pixel labels and instance masks. It is one of the most widely used benchmarks for monocular depth estimation and indoor semantic segmentation.


Changes from the Official Release

  • Converted the official nyu_depth_v2_labeled.mat (MATLAB v7.3) into Parquet (one row per frame), so the Dataset Viewer and datasets work without MATLAB or h5py.
  • Depth (depth, raw_depth): stored as 16-bit PNG in millimetres (round(meters * 1000)), so the maximum quantization error is 0.5 mm. Divide by 1000 to get meters.
  • Labels: label894 is the official label map, unchanged (16-bit PNG). label40 and label13 are derived from it with the standard mappings. Both were checked during the build against Handa's published versions on all 1,449 frames: the 40-class maps match labels40.mat exactly (0 differing pixels), and the 13-class maps match his PNGs 100%.
  • Official 795 / 654 train/test split from splits.mat.
  • Not included: the raw video sequences (~428 GB), accelerometer data, and the toolbox.
  • No images were added, removed, or modified. No labels were changed.

Dataset Structure

Column Content
image RGB, 640x480 PNG
depth filled (in-painted) depth, 16-bit PNG, millimetres
raw_depth raw Kinect depth (with holes = 0), 16-bit PNG, millimetres
label894 original class labels, 0 = unlabeled, 1..894 (names in the official .mat)
label40 40-class labels (0 = unlabeled), Gupta et al. mapping
label13 13-class labels (0 = unlabeled)
instance instance ids within each class
index 1-based frame index in the official .mat
scene scene name, e.g. kitchen_0023
scene_type scene category, e.g. kitchen
raw_rgb_filename matching frame in the raw video dump

Splits: train 795 · test 654 (1,449 total).

Many depth benchmarks evaluate on the valid region [45:471, 41:601] (rows, columns), which excludes the projection border.

Classes

40-class (label40, 1..40): wall, floor, cabinet, bed, chair, sofa, table, door, window, bookshelf, picture, counter, blinds, desk, shelves, curtain, dresser, pillow, mirror, floor mat, clothes, ceiling, books, refridgerator, television, paper, towel, shower curtain, box, whiteboard, person, night stand, toilet, sink, lamp, bathtub, bag, otherstructure, otherfurniture, otherprop

13-class (label13, 1..13): bed, books, ceiling, chair, floor, furniture, objects, picture, sofa, table, tv, wall, window

Usage

import numpy as np
from datasets import load_dataset

ds = load_dataset("dronefreak/NYUv2")
row = ds["train"][0]
rgb = row["image"]
depth_m = np.asarray(row["depth"], dtype=np.float32) / 1000.0   # meters
labels40 = np.asarray(row["label40"])

Dataset Sources

Citation

@inproceedings{silberman2012indoor,
  title     = {Indoor Segmentation and Support Inference from RGBD Images},
  author    = {Silberman, Nathan and Hoiem, Derek and Kohli, Pushmeet and Fergus, Rob},
  booktitle = {ECCV},
  year      = {2012}
}

If you use the 40-class labels, also cite: Gupta, Arbelaez, Malik, Perceptual Organization and Recognition of Indoor Scenes from RGB-D Images, CVPR 2013.


License

The original authors have not specified a license for NYU Depth v2; the project page only requires citation of the paper above. This repository therefore uses license: other and redistributes the data unofficially for research purposes, with attribution. Users are responsible for ensuring their use complies with any terms the original authors may set. Rights holders can request removal by opening a discussion on this repository.

Downloads last month
32

Models trained or fine-tuned on dronefreak/NYUv2

Collection including dronefreak/NYUv2