CHS: Small: Towards Photorealistic Augmented and Virtual Reality Displays

CHS：小型：迈向逼真的增强和虚拟现实显示

基本信息

批准号：
2008464
负责人：
Aswin Sankaranarayanan
金额：
$ 50万
依托单位：
Carnegie-Mellon University
依托单位国家：
美国
项目类别：
Standard Grant
财政年份：
2020
资助国家：
美国
起止时间：
2020-08-01 至 2024-07-31
项目状态：
已结题

来源：
https://www.nsf.gov/awardsearch/showAward?AWD_ID=2008464&HistoricalAwards=false
关键词：
CHS Small Towards Photorealistic Augmented

项目摘要

The holy grail for a 3D display is to produce a scene that, to our eyes, is indistinguishable from reality. Despite significant advances in display technology, simultaneous deception of all perceptual depth cues is still beyond the reach of most displays. As a consequence, existing 3D displays, especially those used in Augmented / Virtual Reality (AR/VR) applications, tend to cause discomfort due to subtle differences between the real and the virtual world. This project aims to develop novel 3D photorealistic display designs that ensure that the light field incident on the eye mimics reality at a significantly higher level of fidelity as compared to existing designs. The development of such a technology stands to impact a broad range of disciplines and applications that go beyond AR/VR systems and applications, such as ophthalmology, and testing of optical systems. Further, as the visual realism of displayed content gets better, they stand to replace standard 2D monitors that are ubiquitous in work and home environments today. These photorealistic displays will have the potential to make AR/VR systems more accessible for visually impaired people. The education and outreach components of this project disseminates display research and demos in middle and high schools in the greater Pittsburgh region via lab visits, and hands-on workshops.To achieve the goal of photorealistic AR/VR displays, this project explores differ approaches for modeling of the virtual world and focuses on the development of three distinct approaches that progressively increase realism at the cost of complexity: (i) simple models in form of depth and texture decomposition; (ii) light field representations that provide a dense sampling of the ray space; (iii) coherent wavefronts using Fourier optics. All the three approaches aim to produce light fields that are dense in spatial and angular resolution even though the underlying physical mechanisms rely on different models for light transport. The project will develop techniques for generating content that is easily adapted to the requirements of the display. The project also explores and characterizes the fundamental properties and features in terms of achievable spatial and angular resolutions, the precision in depiction of occlusion, the size of the eye-box, and finally their limitations.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

3D 显示器的终极目标是产生一个在我们看来与现实无法区分的场景。尽管显示技术取得了重大进步，但同时欺骗所有感知深度线索仍然超出了大多数显示器的能力范围。因此，现有的 3D 显示器，尤其是增强/虚拟现实 (AR/VR) 应用中使用的 3D 显示器，往往会因现实世界和虚拟世界之间的细微差异而引起不适。该项目旨在开发新颖的 3D 逼真显示设计，确保入射到眼睛的光场以比现有设计更高的保真度模拟现实。这种技术的发展将会影响 AR/VR 系统和应用之外的广泛学科和应用，例如眼科和光学系统测试。此外，随着显示内容的视觉真实感变得更好，它们将取代当今工作和家庭环境中普遍存在的标准 2D 显示器。这些逼真的显示器将有可能使视障人士更容易使用 AR/VR 系统。该项目的教育和推广部分通过实验室参观和实践研讨会在大匹兹堡地区的初中和高中传播显示研究和演示。为了实现逼真的 AR/VR 显示的目标，该项目探索了不同的方法虚拟世界的建模，重点关注三种不同方法的开发，这些方法以复杂性为代价逐步提高真实感：（i）深度和纹理分解形式的简单模型； (ii) 提供光线空间密集采样的光场表示； (iii) 使用傅立叶光学的相干波前。所有三种方法的目的都是产生空间和角度分辨率密集的光场，尽管底层的物理机制依赖于不同的光传输模型。该项目将开发生成易于适应显示要求的内容的技术。该项目还探索并描述了可实现的空间和角度分辨率、遮挡描述精度、眼箱大小及其局限性等方面的基本属性和特征。该奖项反映了 NSF 的法定使命，并被视为值得通过使用基金会的智力优点和更广泛的影响审查标准进行评估来支持。

项目成果

期刊论文数量（10）

专著数量（0）

科研奖励数量（0）

会议论文数量（0）

专利数量（0）

HyperReel: High-Fidelity 6-DoF Video with Ray-Conditioned Sampling

HyperReel：采用光线条件采样的高保真 6-DoF 视频

DOI：
10.1109/cvpr52729.2023.01594
发表时间：
2023-01-05
期刊：
2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
影响因子：
0
作者：
Benjamin Attal;Jia;Christian Richardt;Michael Zollhoefer;J. Kopf;Matthew O’Toole;Changil Kim
通讯作者：
Changil Kim

Neural Kaleidoscopic Space Sculpting

神经万花筒空间雕塑

DOI：
发表时间：
2023-01
期刊：
Proceedings IEEE Computer Society Conference on Computer Vision and Pattern Recognition
影响因子：
0
作者：
Ahn, Byeongjoo;De Zeeuw, Michael;Gkioulekas, Ioannis;Sankaranarayanan, Aswin C.
通讯作者：
Sankaranarayanan, Aswin C.

Exponentially-wide etendue displays using a tilting cascade

使用倾斜级联的指数宽集光率显示