SYN-MVMF consists of 600 scenes, each containing 9 RGB multi-view multi-focus images, 9 defocus maps, 9 depth maps, 9 corresponding all-in-focus ground truth images, as well as camera matrices and parameters for each camera pose. We construct two different resolution versions, 1920×1080 and 960×540, to meet the requirements of various tasks.
🔴 Tips: Currently, we provide 5 scenes with a resolution of 480×270 for download. The complete dataset will be released upon publication.
This dataset cannot be used for commercial purposes. For non-commercial research or educational purposes, please inform us for our permission via zhilong_li AT hust DOT edu DOT cn.