- DUTLF-V2 is the updated version of DUTLF-V1, which consists of 4204 light field samples.
- DUTLF-V2 is publicly available in Baiduyun (code: xjur) and Google
Each sample of DUTLF-V2 contains 5 elements, including a RGB image (a), a saliency mask (b), a depth map (c), a stack of 12 focal slices (d) and an array of multi-view images (f).
Our motivation for DUTLF-V2 is to build a versatile dataset in large scale, high diversity and broad coverage. We conduct descriptive statistics on our DUTLF-V2 in terms of the following aspects: Diversity of Salient Objects (a), Challenging Scenes (b and c), Scale of Salient Objects (e) and Distribution of Center Bias (f).
