360°视频分类与数据集
收藏资源简介:
本数据集由德克萨斯大学达拉斯分校和巴西利亚大学合作创建,包含28个基于360°视频分类的场景无剪辑视频。数据集通过60名参与者的视口追踪记录,以及他们对观看视频体验的反馈。数据集旨在研究用户在观看360°视频时的行为模式,特别是在虚拟现实环境中的视口预测和显著性检测。此外,数据集还用于探索360°视频流媒体技术,以及视觉注意和显著性建模。
This dataset was collaboratively developed by The University of Texas at Dallas and the University of Brasília. It comprises 28 uncut videos of scene categories for 360° video classification. The dataset collects viewport tracking data from 60 participants, along with their subjective feedback on their experience of viewing these videos. Its primary goal is to investigate user behavioral patterns during 360° video viewing, with a particular focus on viewport prediction and saliency detection in virtual reality environments. Additionally, this dataset can also be utilized to explore 360° video streaming technologies, as well as visual attention and saliency modeling.

- 1A Taxonomy and Dataset for 360° Videos德克萨斯大学达拉斯分校 · 2019年



