Lytro Illum 光场显着性检测数据集

Lytro Illum 是采用 Lytro-Illum 相机拍摄的光场数据集。该数据集包含 640 个光场及其对应的地面真实显着图。其中与目前公开的光场显着性数据集相比，该新数据集更大，质量更高，包含更多的变化和更多类型的光场输入。这使该数据集适合于训练更深的网络和基准测试。

The Lytro Illum dataset is a light field dataset captured with the Lytro-Illum camera. This dataset contains 640 light fields and their corresponding ground-truth saliency maps. Compared with currently publicly available light field saliency datasets, this new dataset is larger in scale, higher in quality, and covers more variations and more types of light field inputs. This makes the dataset suitable for training deeper neural networks and benchmarking.

搜集汇总

背景与挑战

背景概述

Lytro Illum 光场显着性检测数据集包含640个光场及其对应的地面真实显着图，是目前公开数据集中规模更大、质量更高且变化更丰富的资源。该数据集适用于训练深度网络和进行基准测试。