Object-based motion estimation for highly efficient streaming video

用于高效流视频的基于对象的运动估计

• 批准号: DP190103717
• 负责人: Prof Mark Pickering
• 金额: $ 25万
• 依托单位: The University of New South Wales
• 依托单位国家: 澳大利亚
• 项目类别: Discovery Projects
• 财政年份: 2019
• 资助国家: 澳大利亚
• 起止时间: 2019-01-01 至 2023-10-31
• 项目状态: 已结题
• 来源: https://dataportal.arc.gov.au/RGS/Web/Grants/DP190103717
• 关键词: Object; based; motion; estimation; highly

This project aims to develop motion and shape estimation tools that allow video content to be predicted at any point in time using frame data that is available at other points in time. Applications of this include video coding, frame synthesis for temporal re-sampling. The project expects to develop algorithms for jointly estimating compact motion and shape descriptions of video content. These descriptions will allow quality refinement by progressively adding spatial and temporal detail. It is expected that these compact descriptions will lead to significantly more flexible and efficient transmission of video content over modern communications networks.

该项目旨在开发运动和形状估计工具，允许使用在其他时间点可用的帧数据在任何时间点预测视频内容。其应用包括视频编码、用于时间重采样的帧合成。该项目希望开发用于联合估计视频内容的紧凑运动和形状描述的算法。这些描述将允许通过逐步添加空间和时间细节来改进质量。预期这些紧凑的描述将导致视频内容在现代通信网络上的显著更灵活和有效的传输。