Dataset Viewer
Auto-converted to Parquet Duplicate
Search is not available for this dataset
image
imagewidth (px)
1.02k
2.45k
label
class label
72 classes
0OPID001
0OPID001
0OPID001
0OPID001
1OPID002
1OPID002
1OPID002
1OPID002
2OPID003
2OPID003
2OPID003
2OPID003
3OPID004
3OPID004
3OPID004
3OPID004
4OPID005
4OPID005
4OPID005
4OPID005
5OPID006
5OPID006
5OPID006
5OPID006
6OPID007
6OPID007
6OPID007
6OPID007
7OPID008
7OPID008
7OPID008
7OPID008
8OPID009
8OPID009
8OPID009
8OPID009
9OPID010
9OPID010
9OPID010
9OPID010
10OPID011
10OPID011
10OPID011
10OPID011
11OPID012
11OPID012
11OPID012
11OPID012
12OPID013
12OPID013
12OPID013
12OPID013
13OPID014
13OPID014
13OPID014
13OPID014
14OPID015
14OPID015
14OPID015
14OPID015
15OPID016
15OPID016
15OPID016
15OPID016
16OPID017
16OPID017
16OPID017
16OPID017
17OPID018
17OPID018
17OPID018
17OPID018
18OPID019
18OPID019
18OPID019
18OPID019
19OPID020
19OPID020
19OPID020
19OPID020
20OPID021
20OPID021
20OPID021
20OPID021
21OPID022
21OPID022
21OPID022
21OPID022
22OPID023
22OPID023
22OPID023
22OPID023
23OPID024
23OPID024
23OPID024
23OPID024
24OPID025
24OPID025
24OPID025
24OPID025
End of preview. Expand in Data Studio

OPID

This release contains the OPID images from the local Linear_RGB_Polar RLP collection, with the original training and test split preserved.

Structure

train/OPID001/{0,45,90,135}.png
...
test/OPID071/{0,45,90,135}.png
...

Each scene contains four PNG images named for polarization angles in degrees: 0, 45, 90, and 135. Files are uploaded without image conversion.

All 320 images are 8-bit RGB PNG files. Image dimensions are 1024 x 1024 (264 images), 2048 x 2048 (48 images), or 2448 x 2048 (8 images).

Split Scenes PNG files Bytes
train 72 288 556561531
test 8 32 35695840

Training scenes: OPID001 through OPID070, OPID079, and OPID080. Test scenes: OPID071 through OPID078.

License: CC-BY-4.0.

The accompanying manifest.json records each image's repository path, byte size, and SHA-256 checksum.

Citation

If you use this dataset in your research, please cite:

@article{luo2026updem,
  title   = {UP-Dem: Deep Unrolling Convolutional Sparse Coding for Color Polarization Image Demosaicking},
  author  = {Luo, Yidong and Wu, Caiyun and Li, Chenggong and Wang, Ping and Yuan, Xin and Yang, Kailun and Zhang, Junchao},
  journal = {IEEE Transactions on Image Processing},
  year    = {2026},
  volume  = {35},
  number  = {},
  pages   = {10152-10166},
  doi     = {10.1109/TIP.2026.3730838}
}
Downloads last month
96