CSR: Small: Collaborative Research: EUReCa: Enabling Untethered VR/AR System via Human-centric Graphic Computing and Distributed Data Processing

CSR：小型：协作研究：EUReCa：通过以人为中心的图形计算和分布式数据处理实现不受束缚的 VR/AR 系统

• 批准号: 1717657
• 负责人: Yiran Chen
• 金额: $ 25万
• 依托单位: Duke University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-09-01 至 2020-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1717657&HistoricalAwards=false
• 关键词: CSR; Small; Collaborative; Research; EUReCa

Virtual Reality (VR) and Augmented Reality (AR) devices, especially their mobile versions are newly emergent technologies. However, a major challenge that VR/AR technologies faces is the gap between the increasing needs for graphic and data processing and the limited computing capability of the mobile hardware. Two researchers from GMU and Duke form a team to develop an innovative VR/AR system, namely, "EUReCa", which tackles the challenge of human-centric graphic processing and distributed data processing. This research studies VR/AR system design using standard workloads to understand the computation source utilization that leads to development of usage model, and reduces the computation loads via task allocation, thereby enhancing the efficiency and scalability of computation. The research outcomes will benefit both research and industry at large by integrating the innovations of human interaction and advanced data processing technologies. The education plan enhances existing curricula and pedagogy by integrating interdisciplinary modules on computer graphic, embedded systems, and machine learning with newly developed teaching practices, and gives special attention to women and underrepresented minority groups. The project performs three tasks. Task 1 models the computation resource utilization of VR/AR systems by considering the system configuration and dynamics of user operations. Task 2 explores efficient human-centric graphic rendering framework for reducing computation loads of VR/AR systems. Task 3 exploits novel schemes to enhance computation efficiency via balancing the computation loads and data allocations in graphic rendering and deep neural network (DNN) applications. The techniques will be evaluated on mobile devices.

虚拟现实（VR）和增强现实（AR）设备，尤其是其移动版本是新兴的技术。但是，VR/AR Technologies面临的一个重大挑战是图形处理和数据处理需求不断增长的差距以及移动硬件的计算能力有限。 GMU和Duke的两名研究人员组成了一个团队，以开发创新的VR/AR系统，即“ Eureca”，该系统应对以人为中心的图形处理和分布式数据处理的挑战。这项研究研究了使用标准工作负载的VR/AR系统设计，以了解导致使用模型开发的计算源利用，并通过任务分配减少计算负载，从而提高计算的效率和可扩展性。研究结果将通过整合人类互动和高级数据处理技术的创新来使研究和行业受益。该教育计划通过在计算机图形，嵌入式系统上整合跨学科模块，并与新开发的教学实践相结合，并特别关注女性和代表性不足的少数群体，从而增强了现有的课程和教学法。该项目执行三个任务。任务1通过考虑系统配置和用户操作的动态来对VR/AR系统的计算资源利用进行建模。任务2探讨了有效的以人为中心的图形渲染框架，以减少VR/AR系统的计算负载。任务3利用新的方案来通过平衡图形渲染和深神经网络（DNN）应用中的计算负载和数据分配来提高计算效率。这些技术将在移动设备上进行评估。

项目成果

AdverQuil: an efficient adversarial detection and alleviation technique for black-box neuromorphic computing systems

• DOI: 10.1145/3287624.3288753
• 发表时间: 2019-01
• 期刊: Proceedings of the 24th Asia and South Pacific Design Automation Conference
• 作者: Hsin-Pai Cheng;Juncheng Shen;Huanrui Yang;Qing Wu;Hai Helen Li;Yiran Chen

Toward Efficient and Adaptive Design of Video Detection System with Deep Neural Networks

• DOI: 10.1145/3484946
• 发表时间: 2022-05
• 期刊: ACM Transactions on Embedded Computing Systems (TECS)
• 作者: Jiachen Mao;Qing Yang;Ang Li;Kent W. Nixon;H. Li;Yiran Chen

MobiEye: An Efficient Cloud-based Video Detection System for Real-Time Mobile Applications

• DOI: 10.1145/3316781.3317865
• 发表时间: 2019-06
• 期刊: 2019 56th ACM/IEEE Design Automation Conference (DAC)
• 作者: Jiachen Mao;Qing Yang;Ang Li;H. Li;Yiran Chen

MVStylizer: an efficient edge-assisted video photorealistic style transfer system for mobile phones

• DOI: 10.1145/3397166.3409140
• 发表时间: 2020-05
• 期刊: Proceedings of the Twenty-First International Symposium on Theory, Algorithmic Foundations, and Protocol Design for Mobile Networks and Mobile Computing
• 作者: Ang Li;Chunpeng Wu;Yiran Chen;Bin Ni

Running sparse and low-precision neural network: When algorithm meets hardware

• DOI: 10.1109/aspdac.2018.8297378
• 发表时间: 2018-01
• 期刊: 2018 23rd Asia and South Pacific Design Automation Conference (ASP-DAC)
• 作者: Bing Li;W. Wen;Jiachen Mao;Sicheng Li;Yiran Chen;H. Li