--- license: cc-by-4.0 task_categories: - image-to-image tags: - polarization - linear-rgb - opid size_categories: - n<1K --- # OPID This release contains the OPID images from the local Linear_RGB_Polar RLP collection, with the original training and test split preserved. ## Structure ```text 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: ```bibtex @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} } ```