DUT-LF-V2

DUT-LF-V2是由大连理工大学信息与通信工程学院创建的大型光场数据集，包含4204个样本，分为102个子类。该数据集涵盖多种室内外场景，包括RGB图像、深度图、焦点堆栈和多视角图像，适用于RGB、RGB-D和光场显著性检测。创建过程涉及数据收集、标注和分割，旨在解决光场数据在规模、类别和元素类型方面的不足，以及高维数据处理和模型泛化的问题。

DUT-LF-V2 is a large-scale light field dataset developed by the School of Information and Communication Engineering of Dalian University of Technology. It contains 4204 samples categorized into 102 subclasses, covering diverse indoor and outdoor scenarios. The dataset includes RGB images, depth maps, focus stacks and multi-view images, and is applicable to RGB, RGB-D and light field saliency detection tasks. The construction of this dataset involves data collection, annotation and segmentation, aiming to address the deficiencies of existing light field datasets in terms of scale, category diversity and element types, as well as the challenges in high-dimensional data processing and model generalization.