首页> 外文期刊>Multimedia Systems >Extracting-mapping scheme for the dynamic details in fluid re- simulations from videos
【24h】

Extracting-mapping scheme for the dynamic details in fluid re- simulations from videos

机译:视频中流体模拟中动态细节的提取映射方案

获取原文
获取原文并翻译 | 示例
           

摘要

Reconstructing a 3D fluid simulation (re-simulation) from video has practical significance. In this study, we address the realistic problem of fluid re-simulation in an inverse project. State-of-the-art studies on the inverse problem of fluid simulation have mainly focused on dimensionality reduction for acquiring better time performance, but the realistic aspect has rarely been investigated. This paper presents an extracting-mapping scheme to tightly couple fluid re-simulation with enhanced physically driven data for realistic high-quality dynamic detail. We make a full use of the details extracted from recovered physically driven data to improve the coarse and unrealistic re-simulation from fluid auto-advection. Two schemes are discussed. The first scheme is the density block method (DBM). In this method, a density block database is constructed from the prepartitioned and sorted density blocks, and then, some selected density blocks with dynamic details are selected from the preconstructed database and coupled coherently into the physically driven data to enhance the detail in every auto-advection cycle. The second method is the density spectrum block method (DSBM) in the frequency domain. Using the DSBM and DBM, realistic effects are achieved by extensive quantitative and qualitative evaluation via re-simulation tests driven by the recovered physical data from ground truth fluid video. Both approaches outperform previous auto-advection re-simulation schemes in terms of the rich detail under several challenging scenarios and low-level hardware conditions.
机译:从视频重建3D流体模拟(重新模拟)具有实际意义。在这项研究中,我们解决了逆向工程中流体重新模拟的现实问题。关于流体模拟反问题的最新研究主要集中在降维以获取更好的时间性能上,但很少研究现实方面。本文提出了一种提取映射方案,以将流体重新模拟与增强的物理驱动数据紧密耦合,以实现逼真的高质量动态细节。我们充分利用了从恢复的物理驱动数据中提取的细节,以改善流体自动对流的粗略和不现实的重新模拟。讨论了两种方案。第一种方案是密度块方法(DBM)。在这种方法中,根据预先划分和排序的密度块构建密度块数据库,然后,从预先构建的数据库中选择一些具有动态细节的密度块,并将其连贯地耦合到物理驱动的数据中,以增强每个自动选样中的细节。平流周期。第二种方法是频域中的密度谱块方法(DSBM)。使用DSBM和DBM,通过从地面真流体视频中恢复的物理数据驱动的重新模拟测试,通过广泛的定量和定性评估,可以实现逼真的效果。在一些具有挑战性的方案和低级别的硬件条件下,这两种方法的丰富细节都优于以前的自动平流重新模拟方案。

著录项

  • 来源
    《Multimedia Systems》 |2019年第4期|371-381|共11页
  • 作者单位

    East China Normal Univ, Sch Comp Sci & Software Engn, Shanghai 200062, Peoples R China;

    East China Normal Univ, Sch Comp Sci & Software Engn, Shanghai 200062, Peoples R China;

    East China Normal Univ, Sch Comp Sci & Software Engn, Shanghai 200062, Peoples R China;

    East China Normal Univ, Sch Comp Sci & Software Engn, Shanghai 200062, Peoples R China;

  • 收录信息 美国《科学引文索引》(SCI);美国《工程索引》(EI);
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

    Fluid; Re-simulation; Dynamic detail; Density;

    机译:流体;重新模拟;动态细节;密度;

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号