Overview of Full Scene Volumetric Video Dataset

The Full Scene Volumetric Video Dataset is the first dataset of volumetric videos depicting the interaction of people in real-life scenarios with external scenes.

• It contains 500GB+ of volumetric video data, added up to 10000+ frames.
• The data sequences are captured using Azure Kinect RGBD cameras and has been finely post-processed.
• The dataset has a rich diversity of human-human and human-scene interactions under 4 dimensions.

Our dataset can support many research tasks in volumetric video compression, streaming and other areas, for example:

• Virtual & Augmented Reality
• Social Interaction
• Healthcare
• AI Systems

Other Dataset

Here are disparities between our dataset and exist several datasets that provide head movement tracks of users.
Our dataset is the first that contains surrounding environment and human-environment interaction.

Table 1: Existing datasets.

Publications

The following paper describes in depth: from the data collection and processing to detailed statistics about the data. Kindly remind that the usage of Full Scene Volumetric Video Dataset should be based on a proper cite of this paper.

Kaiyuan Hu, Yili Jin, Haowen Yang, Junhua Liu, Fangxin Wang SSE and FNii, The Chinese University of Hong Kong, Shenzhen Versee Inc. Peng Cheng Laborator FSVVD: A Dataset of Full Scene Volumetric Video 2023.

Volumetric Videos

Here is a index of content of the dataset

Table 2: Index of Content