Dataset Viewer
Auto-converted to Parquet Duplicate
scene_id
stringlengths
12
12
hsi_path
stringlengths
27
34
split
stringclasses
3 values
used_in_repository_train_list
bool
2 classes
hdf5_readable
bool
2 classes
height
float64
482
482
⌀
width
float64
512
512
⌀
num_bands
float64
31
31
⌀
hdf5_shape
stringclasses
1 value
dtype
stringclasses
1 value
wavelength_start_nm
float64
400
400
⌀
wavelength_end_nm
float64
700
700
⌀
wavelength_step_nm
float64
10
10
⌀
norm_factor
float64
0.18
20.7
⌀
minimum
float64
-0
0.09
⌀
maximum
float64
1
1
⌀
nonfinite_values
float64
0
0
⌀
zero_values
float64
0
65
⌀
bytes
int64
19.1M
24.2M
sha256
stringlengths
64
64
archive_crc32
stringlengths
8
8
source_issue
stringclasses
1 value
ARAD_1K_0001
Train_Spec/ARAD_1K_0001.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.496855
0.008803
1
0
0
21,896,837
d75d3689827d69fdb600ef49eceec4692615db5f0bb8882b41cc3cc0f29a139f
96de6b79
null
ARAD_1K_0002
Train_Spec/ARAD_1K_0002.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.504815
0.004243
1
0
0
22,920,548
5ee379c0574ac7a225cdd6cd39f942441f48d3401f4d1abe0e26a664ce8f1af3
74eeea50
null
ARAD_1K_0003
Train_Spec/ARAD_1K_0003.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.272186
0.018211
1
0
0
21,218,912
0729ab86549ca910803e2d0dcdc76c10087f9a0f4efbb252e08b564bd3a9a741
a61c40eb
null
ARAD_1K_0004
Train_Spec/ARAD_1K_0004.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.659921
0.006252
1
0
0
22,377,073
abe2de20dc3b76e1a1701292e7ad1774aef643a88366084eef58d92f7e280a6d
1fd4ef16
null
ARAD_1K_0005
Train_Spec/ARAD_1K_0005.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.652874
0.002675
1
0
0
21,042,006
6fd2f928230e076be0d2c707b657a5e6cdf84108cd53d86aab91450d8ce4222b
073fefe2
null
ARAD_1K_0006
Train_Spec/ARAD_1K_0006.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.482583
0.006501
1
0
0
22,078,636
32d2e4ff2732cd16bf3a9103592c9af304598307dac6967990bd25f00b47d9f7
fff2008e
null
ARAD_1K_0007
Train_Spec/ARAD_1K_0007.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.562661
0.008995
1
0
0
23,345,013
70a1d92e7b577fcb436629ba85233ee53279e1847f91ca2abc0bc1acc844b58d
c57eae89
null
ARAD_1K_0008
Train_Spec/ARAD_1K_0008.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.282711
0.022596
1
0
0
21,515,160
888017dabf0180a9f271972d29fbaa8fe1fe25dc4cb7c035a3ed0f859fac1815
4d3fc1d0
null
ARAD_1K_0009
Train_Spec/ARAD_1K_0009.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.482189
0.002405
1
0
0
22,704,229
a728087bc4b7fa1ef7e9b00e0bd6597f5ebc46ef3a65785247bc21e6a8313bc5
99073982
null
ARAD_1K_0010
Train_Spec/ARAD_1K_0010.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.654874
0.006618
1
0
0
22,035,658
8370f939305c5013fe9fb6c4d3d2e02d25c074d5c218abc74c40bf0f95b9b052
fdef2f9a
null
ARAD_1K_0011
Train_Spec/ARAD_1K_0011.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.823911
0.007382
1
0
0
23,405,030
0c1ac971c15d3a67fe352adc243dc9788c84603db7854b06e55d93d4798871f5
c41233ad
null
ARAD_1K_0012
Train_Spec/ARAD_1K_0012.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.945919
0.004442
1
0
0
23,665,263
ea39dca6d5bd3866d4f09f65fe724bdfc012bc300103497f6ea1993e40cfe84b
2f3e5638
null
ARAD_1K_0013
Train_Spec/ARAD_1K_0013.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.112397
0.001825
1
0
0
23,691,329
e6e821070f00e2873b587714dafcdc6274d9b66f1af9b8c9039381030196d602
52ccdaf8
null
ARAD_1K_0014
Train_Spec/ARAD_1K_0014.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.323876
0.001253
1
0
0
23,752,008
c1fb13e2649b30311078583101ee66e65399462d3ab0f3a9012eabf7aebb47e1
11851d28
null
ARAD_1K_0015
Train_Spec/ARAD_1K_0015.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.27172
0.015946
1
0
0
20,590,534
9ed673cf527a461af78dc68e0804c6356bb0513fdb2259b2e2a8be45bdda4716
c2929268
null
ARAD_1K_0016
Train_Spec/ARAD_1K_0016.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.300583
0.016414
1
0
0
20,464,654
544c0716c5783bddbdc33282aa111a17d5648da386b696472ede0687a9f00922
373baa0c
null
ARAD_1K_0017
Train_Spec/ARAD_1K_0017.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.466738
0.012793
1
0
0
21,437,311
74647fad6e446619143c7961270a46f461c27bc9bdb4d7bd8b837dc36209201b
5c3dea0a
null
ARAD_1K_0018
Train_Spec/ARAD_1K_0018.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.397413
0.011526
1
0
0
21,576,435
262b2d41ca6ad0bb8983960cca5159507f9939a9e930346cb22fa84973caff74
3f671702
null
ARAD_1K_0019
Train_Spec/ARAD_1K_0019.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.353187
0.011172
1
0
0
21,386,655
a1308ec6f1da75a2677b58af7078a2e54d577b8246fd642164c4d97f586087f6
0dd90d8f
null
ARAD_1K_0020
Train_Spec/ARAD_1K_0020.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.631861
0.007964
1
0
0
22,981,374
613b095343688c888a69af4e36d971529f3643a828355219632107f5344a4681
3bec3de8
null
ARAD_1K_0021
Train_Spec/ARAD_1K_0021.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.997172
0.002228
1
0
0
23,270,795
533dd7fcc6ac1bfe540929a3aaf3511c9b83d0bd87ede504449f44af0ed80647
d3736451
null
ARAD_1K_0022
Train_Spec/ARAD_1K_0022.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.159572
0.002707
1
0
0
23,277,711
1e602fcd1e46c552d1b156f65c927bcbd2eba3659975aa0a15182dbe2d590000
c8a076f1
null
ARAD_1K_0023
Train_Spec/ARAD_1K_0023.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.624826
0.005884
1
0
0
21,003,269
220625545199fe6da0abeaadc2dbf0750c20e1e12cfadcb2f21d73aa297aec3c
4594633d
null
ARAD_1K_0024
Train_Spec/ARAD_1K_0024.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.320292
0.01252
1
0
0
21,319,051
4283b3e4542d0b7e031053c076bd9fcd77cc1e4ca61a2c0732f9eca82a0a147e
f8f91ff4
null
ARAD_1K_0025
Train_Spec/ARAD_1K_0025.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.291889
0.011796
1
0
0
21,103,154
82bac04e1aa0920e55b54b38deeda2e86f5fa746ab8328802a9800d1923831be
aa5de283
null
ARAD_1K_0026
Train_Spec/ARAD_1K_0026.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.249038
0.022867
1
0
0
20,006,086
4dd73aec98629a600e4740bfa6614a577e5fe4af154862b72955dff80ec822bb
be0810a3
null
ARAD_1K_0027
Train_Spec/ARAD_1K_0027.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.302577
0.013956
1
0
0
20,397,982
ec63a5a3529b65d677569fbf5448dc47a4f0ccbd2fe0cef1abb9bf6c0dbbf046
982b5cd3
null
ARAD_1K_0028
Train_Spec/ARAD_1K_0028.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.404717
0.016446
1
0
0
21,273,967
1d5ee74aa2533a81963102c0a9139eb009e35f7cd951eb632bd90e397c147a9d
f5bffa8e
null
ARAD_1K_0029
Train_Spec/ARAD_1K_0029.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.358952
0.021062
1
0
0
20,616,004
43b655f2f173dfd99c4d84a4e6f0ee3fb82abdcc9441b952ba2563bd3f61c1cc
a2db0399
null
ARAD_1K_0030
Train_Spec/ARAD_1K_0030.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.321264
0.010984
1
0
0
21,242,860
b6c8edc4c67455136b09a2f8dacf7086bda0483047436d391d83aa3f92463c75
b8820837
null
ARAD_1K_0031
Train_Spec/ARAD_1K_0031.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.349211
0.009867
1
0
0
20,795,414
bbb14f9d905e1484dacf3a0798e2f3f23baf03d4307e54699e0fc9d5261fd735
cff83946
null
ARAD_1K_0032
Train_Spec/ARAD_1K_0032.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.523265
0.001523
1
0
0
23,723,706
ad3d852a4485411ec7b830326e0faa7114c8559af9f15ec5010c69f8b52fa3a6
2c3c3c40
null
ARAD_1K_0033
Train_Spec/ARAD_1K_0033.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
4.190098
0.000949
1
0
0
23,738,184
1d24cd7d7caf568b918b02f54c04a44fb7bb6ddb634a3a04fb6e7acd367eb508
6f0b79aa
null
ARAD_1K_0034
Train_Spec/ARAD_1K_0034.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.911361
0.001588
1
0
0
22,098,449
d4a76202be09a3cc10ae7ea3cfbd639538625ff58e9b6347b648625a9497d962
06c2bc63
null
ARAD_1K_0035
Train_Spec/ARAD_1K_0035.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.635707
0.002658
1
0
0
22,169,811
4f6aae1ccd48d5534f52c4e4986572f716375fbc4912dfaaef01bf5e0cdac421
d77a5208
null
ARAD_1K_0036
Train_Spec/ARAD_1K_0036.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.762061
0.00218
1
0
0
22,677,052
3a6aaa62135b5c68b7a836fa4b2dc08c1e1e96fc14331fa85a1d3d44dc715364
fdcb68c2
null
ARAD_1K_0037
Train_Spec/ARAD_1K_0037.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.288401
0.00172
1
0
0
22,643,493
8f7b8a9a19b1ad4bbe7483e8f5b4508843f28dbde238d9477890f25c80138c72
56ef40e2
null
ARAD_1K_0038
Train_Spec/ARAD_1K_0038.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.481504
0.006459
1
0
0
21,961,658
482566b476ea6e0aa3939e72781812d3dcebd64e957616f8175712f89ca8e828
23caf4c3
null
ARAD_1K_0039
Train_Spec/ARAD_1K_0039.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.657455
0.004354
1
0
0
22,934,692
9eaedde59b2033d9d7d85041e3837727284f248ef8139e8a07b022b3760f5479
f2e015c2
null
ARAD_1K_0040
Train_Spec/ARAD_1K_0040.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.896499
0.003852
1
0
0
22,529,482
14dbbadc33709af1da1d8818132971ea57cd42f1910daafeabb91756e3b1e6f7
9070b36a
null
ARAD_1K_0041
Train_Spec/ARAD_1K_0041.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.23205
0.004196
1
0
0
23,394,475
4736bb22a16238b9b373ff61eef4eb6bbd3d312108e00cb24614fe40d96052f2
b672c819
null
ARAD_1K_0042
Train_Spec/ARAD_1K_0042.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.660042
0.005399
1
0
0
22,811,423
9b09dddccc8eff28db490501d579dfddf9d320cab7d98c193b58be08e6b8a01d
e351149c
null
ARAD_1K_0043
Train_Spec/ARAD_1K_0043.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
3.821488
0.001316
1
0
0
23,605,512
bf1d96fd23776ddd1a8e69837f8d1512824db13d041e9004699c906fb49e2d7e
1a1703d5
null
ARAD_1K_0044
Train_Spec/ARAD_1K_0044.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.896702
0.004657
1
0
0
23,531,988
baaf955300e172352d42486e80b7134fafde69964b8d5a7603873791ef2f7a05
5ef123e1
null
ARAD_1K_0045
Train_Spec/ARAD_1K_0045.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.505014
0.006447
1
0
0
22,083,036
1ae502257ca90a27598ed8276767ec1fb657d8e29cd5e71a8f47db26298b17c7
e331522b
null
ARAD_1K_0046
Train_Spec/ARAD_1K_0046.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
3.756523
0.001276
1
0
0
21,464,090
0fd933026404d6f296c0350b9fc42bb4af54848125a002ecf556ffac465295ab
a8a72351
null
ARAD_1K_0047
Train_Spec/ARAD_1K_0047.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.803632
0.003214
1
0
0
21,797,143
8bd9f534053ce99ae739b75099202b6d95a861d51fd6cc7debd66344e428f5e7
0f600807
null
ARAD_1K_0048
Train_Spec/ARAD_1K_0048.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.768678
0.003425
1
0
0
21,697,086
3fa4af9db560afe70a59662799d2f4622a17ff688cde1a3473a62ca56a6cae31
6f35923c
null
ARAD_1K_0049
Train_Spec/ARAD_1K_0049.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.449803
0.001605
1
0
0
22,368,429
fb6ad8e9d87a4d71416d2cf278177a8097f5a76ee73fd381b443cd44cc6184e4
782598e8
null
ARAD_1K_0050
Train_Spec/ARAD_1K_0050.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.321973
0.000534
1
0
0
21,577,137
db5daa2fa79e39c5d569711ef4e6733632297157fcd32615ce4dbe66260dba12
92dde9d3
null
ARAD_1K_0051
Train_Spec/ARAD_1K_0051.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
4.656016
0.001129
1
0
0
23,711,818
3aa9b9ae2e800281dfee5ebc1ff9ee7b4d964722cb9850f2f0534fbd279ed90a
46404ead
null
ARAD_1K_0052
Train_Spec/ARAD_1K_0052.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.272446
0.002572
1
0
0
23,585,046
0b520094d9dd7c02f0baa53a833a3edeffb3801d03cc5874897c84f59e6b74e3
dbbda43f
null
ARAD_1K_0053
Train_Spec/ARAD_1K_0053.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.15268
0.004337
1
0
0
22,359,219
e9e3b7fb9dc714ba6aa5cf340fa16b51416b3484960fe79bfb617e103b136c6b
a1393ab9
null
ARAD_1K_0054
Train_Spec/ARAD_1K_0054.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.415727
0.010176
1
0
0
20,306,168
501c7c0f4e52f6239e93485735d967a47ebc24f16124fed50527bd977488f3df
8ad32d03
null
ARAD_1K_0055
Train_Spec/ARAD_1K_0055.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.383883
0.012223
1
0
0
21,515,334
5e9b4a0e1eae733b854799df03cbc0ab85d97f5a7c6e0723844aa7b98e8a3f30
26d62e98
null
ARAD_1K_0056
Train_Spec/ARAD_1K_0056.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.943833
0.007554
1
0
0
23,255,687
e9ca2f2b5ad149ee5ab4197d1c6262e6cf0dc392f2a1716884759034f92db363
eafa1a5e
null
ARAD_1K_0057
Train_Spec/ARAD_1K_0057.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.207929
0.000412
1
0
0
21,926,461
bb425bfd0b3f8c70ce0f96585d18a1c9ef470b8711f38eda9ad4f437b717f1d1
314b9075
null
ARAD_1K_0058
Train_Spec/ARAD_1K_0058.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.945833
0.002106
1
0
0
23,366,847
737c4902d5a84c9b624a42adf04978f4ea2c3d4db86c3a587051685ed00c804e
73d45c58
null
ARAD_1K_0059
Train_Spec/ARAD_1K_0059.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.693097
0.002953
1
0
0
22,092,323
27fbd463e72643deac7c1780f0e5d9cf5bb94e1b4c5e5812f926c45acf4ad991
43cf085f
null
ARAD_1K_0060
Train_Spec/ARAD_1K_0060.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.677725
0.004167
1
0
0
22,904,404
1601190bedee97941312ce4a4ef199d131bc35a243482dc6e783db1f3bea82a4
27f07003
null
ARAD_1K_0061
Train_Spec/ARAD_1K_0061.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.754682
0.000479
1
0
0
23,357,424
b4f984265c94d51f44734a49cc32dfe6eee49f0990682bc923c0a21f2f6d0e62
14cb0535
null
ARAD_1K_0062
Train_Spec/ARAD_1K_0062.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.442013
0.001203
1
0
0
22,926,196
10d02591607a8f8214a270e57e515d49d06e1a29d300c006aff748d0d9f67d3b
f92e54b2
null
ARAD_1K_0063
Train_Spec/ARAD_1K_0063.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.48342
0.005048
1
0
0
21,226,262
8f466f042c4b96257d1cea98e477ea18ed896e4065e38e148dfa0bd2bc0249b9
dd803ebd
null
ARAD_1K_0064
Train_Spec/ARAD_1K_0064.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.43267
0.006237
1
0
0
22,881,999
b98a6cbc9104704bc2fd0335a9537c7ce8b7e325c1c6d255566142f10b5efcbc
8c16c9fb
null
ARAD_1K_0065
Train_Spec/ARAD_1K_0065.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.687075
0.004673
1
0
0
23,188,340
3472e8e756da7574783bdae70a31bcce873bab85eb6589e13ec61570c43f4a5e
39365401
null
ARAD_1K_0066
Train_Spec/ARAD_1K_0066.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.932073
0.00235
1
0
0
22,176,344
26d6dd1e404bde91874f591bcd6fa3fc495429fbd3586a8575fda400e7faa6a5
92e65a83
null
ARAD_1K_0067
Train_Spec/ARAD_1K_0067.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.453191
0.000725
1
0
0
24,188,094
8a460aebcfeda6826568f79d88332930ae92857876246f6a37e5c2ad20c7d69f
1180abc0
null
ARAD_1K_0068
Train_Spec/ARAD_1K_0068.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.733371
0.002487
1
0
0
23,632,996
a116f85bc7d9e75d8249863b457181ff0ce0b04ee3977496cbf4a174e79308fc
227900b4
null
ARAD_1K_0069
Train_Spec/ARAD_1K_0069.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.077481
0.001803
1
0
0
22,705,186
a2f3d6c02c1b042148b3f3103bb053c6323fc148c5fbbdbf431be2ed86b2a47a
652e5858
null
ARAD_1K_0070
Train_Spec/ARAD_1K_0070.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.317313
0.001718
1
0
0
22,765,378
d7d84a5f5c13bb05f696f05dcef7e903f288bc89f5f3aa426d25ec428734e6f0
654cc359
null
ARAD_1K_0071
Train_Spec/ARAD_1K_0071.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
7.07609
0.000766
1
0
0
22,506,754
e474aad44bd1e12a5d8bc7644d8b4bbfac4a3cffba3b0440674c36d3b576fa8b
bd8c7641
null
ARAD_1K_0072
Train_Spec/ARAD_1K_0072.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.018942
0.00188
1
0
0
22,047,959
794a6aa466c66a183a9377cd9bc41dafa24f39e5c93df966b412d00ab95aa46c
ae206705
null
ARAD_1K_0073
Train_Spec/ARAD_1K_0073.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
3.37023
0.000981
1
0
0
22,624,706
d4bb3b41e3cb17a83c564f76c5986d0d1c23d1b6c5757729abcbaedc3f611cab
dd7b8742
null
ARAD_1K_0074
Train_Spec/ARAD_1K_0074.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.839017
0.00156
1
0
0
22,173,860
5d049f150920bd2f84811e898038f57dedd1375079b764a414c460be9dc3f11c
bdf5a879
null
ARAD_1K_0075
Train_Spec/ARAD_1K_0075.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
3.498978
0.001066
1
0
0
22,553,710
8de9d0142255d708f6d5a8e06a6e118fc944a69403cec876c85c7ab4b71565c9
541a81ca
null
ARAD_1K_0076
Train_Spec/ARAD_1K_0076.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.499425
0.00128
1
0
0
22,948,930
da6c1afe1b2628af406c89eb3b69101a266439e5945982acd315660001d3c27e
ae8ecee5
null
ARAD_1K_0077
Train_Spec/ARAD_1K_0077.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.914644
0.001582
1
0
0
22,954,695
c926d8a144776fc8e575410cbc3f1cef72ee6a70972d812d33c68d0b9ad8596e
172b7a41
null
ARAD_1K_0078
Train_Spec/ARAD_1K_0078.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.469189
0.005281
1
0
0
19,067,563
b056b665496c753f98b6519d82b93719ecda1b39ae6966d5a5b1fb8fb8cf9fce
30789c36
null
ARAD_1K_0079
Train_Spec/ARAD_1K_0079.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.450746
0.001287
1
0
0
22,673,126
4b3e7747202d5febebe9765ff13d0715ffeb9306cdbce86f4dced67e8bc67afe
824f0a11
null
ARAD_1K_0080
Train_Spec/ARAD_1K_0080.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.681809
0.001805
1
0
0
23,221,418
73758878e25c07060bb879646a2847c41f7399dbc315a21c59b58d8aaea2e05f
2dfddc81
null
ARAD_1K_0081
Train_Spec/ARAD_1K_0081.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
3.300184
0.000918
1
0
0
22,993,840
5084518acdb2631ad3160a444fcf92b1a83969084a19fcf49662dd1250576db9
97700cf9
null
ARAD_1K_0082
Train_Spec/ARAD_1K_0082.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.408118
0.001096
1
0
0
23,223,763
1b5ca528513f41c739c1fb3a0895fce24e689d6db3d3e4b278ff263953ba4d25
40e6499e
null
ARAD_1K_0083
Train_Spec/ARAD_1K_0083.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.339855
0.001136
1
0
0
23,362,181
8c872a92dc1c372ade040ecc4bb3affba63dd850ed6471f4a9182b5da00022df
5291ff0a
null
ARAD_1K_0084
Train_Spec/ARAD_1K_0084.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.423603
0.001688
1
0
0
19,864,417
630c8842f3e74a11bdad53b97428bce337cf413133cd2996e804516cdfaea641
260ba6ca
null
ARAD_1K_0085
Train_Spec/ARAD_1K_0085.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.375022
0.008126
1
0
0
19,526,508
2cd0850d320aef90aa1f0d754bbf9025aad8c912e5e55297a91141db4b365c5f
388a84dd
null
ARAD_1K_0086
Train_Spec/ARAD_1K_0086.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.079265
0.002603
1
0
0
23,072,454
32cff57c1d9d2f47e2fa6dc79904882a63254dad43bac56a12dda4e561d087eb
a34eaf62
null
ARAD_1K_0087
Train_Spec/ARAD_1K_0087.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.562948
0.0014
1
0
0
21,463,689
6a1f52f6346144e40983d105ed9a9afb63874f65a1dd4c71cee4459b693d6371
21239661
null
ARAD_1K_0088
Train_Spec/ARAD_1K_0088.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.616241
0.001643
1
0
0
21,998,240
efce92e3a5a2005365b92525d6bfe21918990ed1561cf939c6481e1d3afe0337
dfa08b11
null
ARAD_1K_0089
Train_Spec/ARAD_1K_0089.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.113449
0.001126
1
0
0
23,454,700
5511b678f0d504a1ec0583536b48a26ef885041a1ae0cc005251177a17dbd7fd
8e652d98
null
ARAD_1K_0090
Train_Spec/ARAD_1K_0090.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.1509
0.004638
1
0
0
22,681,665
5affdece00910aae82b7c6feaf970d68842029239f7146eea9e2a0470cbeee78
ce99ca26
null
ARAD_1K_0091
Train_Spec/ARAD_1K_0091.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.776379
0.002929
1
0
0
23,409,708
3ef13b9022e57a77093ab4e2b782509b80ae35274f75022fa62c898017830962
e5fb3142
null
ARAD_1K_0092
Train_Spec/ARAD_1K_0092.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.311997
0.007754
1
0
0
22,968,359
1c5f191d804458554a2130813b7dd2e4892270f3311d1244a98f6358b411d34b
1bb67621
null
ARAD_1K_0093
Train_Spec/ARAD_1K_0093.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.091361
0.00137
1
0
0
20,666,605
c823c0ec59eab707bf265591034253f644584e111556488586c28a7b562517fb
801397ec
null
ARAD_1K_0094
Train_Spec/ARAD_1K_0094.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.215997
0.001965
1
0
0
23,447,960
b0e47c6e4cdf0e165efd3acf65b6f7b8bd98edf13f15df8f48cfdaf3a19ee8a1
8a37f1f0
null
ARAD_1K_0095
Train_Spec/ARAD_1K_0095.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
3.602203
0.000652
1
0
0
23,402,131
b2c4cc2e391e37496d46f394a6f52ec668b13fdbca192978dd636505f08edfde
33fed69c
null
ARAD_1K_0096
Train_Spec/ARAD_1K_0096.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.316029
0.000629
1
0
0
22,976,382
fe3186ace55f0484cb1698583f0b393e3f1df8c04a72a0bdd31256d3f325c58a
f2842e32
null
ARAD_1K_0097
Train_Spec/ARAD_1K_0097.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
3.403515
0.001096
1
0
0
23,362,837
63829cfcd54596521f3eacf7872a8f6f16d2c197f716de895716c4fd9a4626b8
ef62678a
null
ARAD_1K_0098
Train_Spec/ARAD_1K_0098.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
1.87535
0.001192
1
0
0
23,270,625
a5ec668116414301954dee668a097998f3f76d9941d8cefc02f5568d6fe47744
f38b3475
null
ARAD_1K_0099
Train_Spec/ARAD_1K_0099.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
0.466483
0.005002
1
0
0
21,751,911
bf3c15e9302cd8bc332687f356c0718f71d79ec7503aeb4e67545502ccb48c5f
48777a5b
null
ARAD_1K_0100
Train_Spec/ARAD_1K_0100.mat
train
true
true
482
512
31
31x512x482
float32
400
700
10
2.448549
0.002194
1
0
0
20,946,226
877b8b1246213389f2cb25e2f3d125a81f6ace137dc3b229356cbc4404fd288a
2ea99525
null
End of preview. Expand in Data Studio

MST++ · ARAD_1K Training Spectral Images

Training hyperspectral data for RGB-to-HSI spectral reconstruction · NTIRE 2022

Code · Models · Online demo · MST++ paper · Dataset paper

Dataset description

This repository curates the 897 training hyperspectral images selected by the MST++ toolbox from the original 900-file ARAD_1K training archive. The source is linjing7/Train_spectral. All 900 original MAT files are preserved byte for byte: the 897 selected cubes are in Train_Spec/, two readable scenes outside the supplied training list are in additional_scenes/, and one already-truncated source file is retained in source_issues/ for provenance.

ARAD_1K was introduced by Boaz Arad et al. This repository distributes the training spectral archive; it does not introduce a new dataset. The associated MST++ model repository provides MST++ and ten comparison methods from the spectral reconstruction toolbox.

Data at a glance

Property Value
Original archive files 900
Verified training cubes in Train_Spec/ 897
Additional readable cubes 2, in additional_scenes/
Truncated source file 1, in source_issues/; excluded from training
Scene IDs ARAD_1K_0001–ARAD_1K_0900
Spatial resolution 482 × 512
Spectral bands 31; 400–700 nm in 10 nm steps
File format MATLAB 7.3 / HDF5 (.mat)
cube datatype float32
HDF5 array shape in h5py (31, 512, 482)
PyTorch channel-first shape after the repository transpose (31, 482, 512)
MATLAB / channel-last shape (482, 512, 31)
Other retained variables bands, norm_factor
Training-list entries supplied by the MST++ repository 897
Validation spectral images Not included
Paired RGB images Not included

The supplied split_txt/train_list.txt matches all 897 files in Train_Spec/. The three original files outside that list are retained separately:

Original scene Location Observed source-data condition
ARAD_1K_0314 source_issues/ The source archive member is truncated; HDF5 cannot open it.
ARAD_1K_0340 additional_scenes/ Readable; 22 zero values and some negative values.
ARAD_1K_0531 additional_scenes/ Readable; 65 zero values and some negative values.

The 897 selected training cubes are readable, finite, and strictly positive. Zeros make the original MRAE loss's division by the target undefined, so keep the provided split for training. No cube is repaired, clipped, or rewritten. metadata.csv describes the 897 training cubes; archive_manifest.csv records all 900 originals, their relocated paths, checksums, and source issues.

The 50 validation scenes (ARAD_1K_0901–ARAD_1K_0950) are absent from this archive. Original validation and test RGB data, training RGB data, and validation ground-truth cubes must be obtained separately following the project's data preparation instructions.

Ground-truth spectral previews

Wavelength-colored previews of training scene ARAD_1K_0001. These are visualizations of the provided ground-truth cube, with no model inference.

480 nm 520 nm 580 nm 660 nm
480 nm 520 nm 580 nm 660 nm

Download

Run from the MST-plus-plus repository root:

hf download CaiYuanhao/MST_pp --repo-type dataset \
  --include "Train_Spec/*.mat" "split_txt/train_list.txt" "metadata.csv" "dataset_manifest.json" \
  --local-dir dataset

To download one cube:

hf download CaiYuanhao/MST_pp Train_Spec/ARAD_1K_0001.mat \
  --repo-type dataset --local-dir dataset

The MAT files are already extracted; no ZIP extraction is required. The command above downloads the 897 verified training cubes. Download the entire repository without --include to also retain additional scenes, the truncated source member, previews, and the complete archive manifest.

Read a hyperspectral cube

import h5py
import numpy as np

with h5py.File("dataset/Train_Spec/ARAD_1K_0001.mat", "r") as mat:
    # Same orientation conversion as train_code/hsi_dataset.py.
    hsi = np.asarray(mat["cube"], dtype=np.float32).transpose(0, 2, 1)
    wavelengths_nm = np.asarray(mat["bands"]).reshape(-1)
    norm_factor = float(np.asarray(mat["norm_factor"]).squeeze())

print(hsi.shape)       # (31, 482, 512)
print(wavelengths_nm)  # [400., 410., ..., 700.]

The project consumes cube as stored. bands and norm_factor are retained source metadata. The cube is not an RGB image; paired training uses a separate RGB image with the same scene ID.

For the file metadata table rather than the spectral arrays, optionally install datasets (pip install datasets) and run:

from datasets import load_dataset

metadata = load_dataset("CaiYuanhao/MST_pp", "metadata", split="train")
print(metadata[0])

The viewer and metadata configuration show one row per verified training cube (897 rows), including its path, dimensions, wavelength range, normalization factor, training-list membership, value statistics, byte size, and SHA256 hash. They do not embed the 31-band arrays into Parquet. The archive_manifest configuration, split archive, additionally exposes all 900 source records.

Provenance and integrity

Source archive: Train_spectral.zip, revision c57ff1dce40b55c47e908b010b744f1382425f11.

dataset_manifest.json records the archive SHA256, source revision, array layout, and validation summary. Every archive member is checked against its ZIP CRC32 after extraction and has a recorded SHA256 hash. All 899 readable cubes are checked for shape, wavelength labels, and finite values. The one truncated member matches its source ZIP CRC32 and is explicitly flagged; it is not presented as a usable training cube. Per-file hashes are available in the metadata tables. No MAT variables or pixel values are rewritten.

Usage and limitations

These data support RGB-to-hyperspectral reconstruction research when combined with the matching RGB inputs. This training-only mirror cannot independently reproduce validation metrics or challenge test rankings. Ground-truth data, model predictions, and display pseudocolors are distinct representations; preview PNGs are not training targets.

Original ARAD_1K / NTIRE dataset terms apply. The mirrored archive does not include a separate data license. The software/model license is not assigned to these dataset files by this mirror.

Citation

Please cite the dataset report and the method used in your experiments:

@inproceedings{Arad_2022_CVPR,
  title={NTIRE 2022 Spectral Recovery Challenge and Data Set},
  author={Arad, Boaz and Timofte, Radu and Yahel, Rony and Morag, Nimrod and Bernat, Amir and others},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
  pages={863--881},
  year={2022}
}

@inproceedings{mst_pp,
  title={MST++: Multi-stage Spectral-wise Transformer for Efficient Spectral Reconstruction},
  author={Yuanhao Cai and Jing Lin and Zudi Lin and Haoqian Wang and Yulun Zhang and Hanspeter Pfister and Radu Timofte and Luc Van Gool},
  booktitle={CVPRW},
  year={2022}
}
Downloads last month
63

Paper for CaiYuanhao/MST_pp