- This dataset consists of DUTLF-MV, DUTLF-FS and DUTLF-Depth.
- This dataset will be expanded to 3000 or so real scenes.
- We are working on it and will make it publicly available soon.
- DUTLF-FS is part of DUTLF, which consists of 1465 real scenes.
- Each scene of this dataset consists of an all-focus image, focal stack and corresponding ground truth.
- Dataset can be downloaded from Google Drive or Baidu Drive.
- Training set: 1000 samples before data augmentation
- Test set: 465 samples