Real-time 3D object/body pose estimation

实时 3D 物体/身体姿势估计

• 批准号: 521969-2017
• 负责人: JanabiSharifi, Farrokh
• 金额: $ 1.82万
• 依托单位: Ryerson University
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2017
• 资助国家: 加拿大
• 起止时间: 2017-01-01 至 2018-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=640089
• 关键词: Real; time; 3D; object; body

With increasing interest in the applications of human and robot interactions, the need for easy robotprogramming and their understanding from their surrounding has become a priority. In particular, there is astrong demand for robotic systems in various fields including telepresence, security and health monitoring thathave become very popular in many industrial and commercial applications. Majority of the previousapproaches, developed many years ago, are designed for an environment under various limitations and are notapplicable to commercial products. Besides they have been mainly devised for laboratory conditions orcontrolled challenges such as Amazon grand challenge.A major challenge of the industry is to develop fast and robust understanding of the objects or bodies in theenvironment seen by the robot and estimate their 3D poses accurately. Crosswing is pursuing the goal toempower family, friends, shoppers, travelers, doctors, clinicians, security and office staff to reach out to othersby coupling leading media devices with semi-autonomous motion systems, through its patented virtualMETMIPR platform. As a result, they plan to meet the challenge of delivering low cost, yet real-world functionality infields spanning from health care to personal security. Development and integration of the proposed poseestimation method will expand the company's market and competitiveness. In addition to the benefits to thecompany, the proposed work will create training and job opportunities for the graduate students andpostdoctoral fellows engaged in this project and/or its continuation.

随着人们对人类和机器人交互应用的兴趣日益浓厚，对简单的机器人编程及其对周围环境的理解的需求已成为当务之急。特别是远程呈现、安全和健康监控等各个领域对机器人系统的需求强劲，这些领域在许多工业和商业应用中都非常受欢迎。以前的大多数方法是多年前开发的，是针对各种限制的环境而设计的，不适用于商业产品。此外，它们主要是为实验室条件或受控挑战（例如亚马逊挑战赛）而设计的。该行业的一个主要挑战是对机器人所看到的环境中的物体或身体进行快速而可靠的理解，并准确估计它们的 3D 姿势。 Crosswing 的目标是通过其获得专利的虚拟METMIPR 平台将领先的媒体设备与半自主运动系统相结合，从而使家人、朋友、购物者、旅行者、医生、临床医生、保安和办公室工作人员能够与他人联系。因此，他们计划迎接挑战，在医疗保健和个人安全等领域提供低成本但真实的功能。所提出的姿态估计方法的开发和集成将扩大公司的市场和竞争力。除了给公司带来好处外，拟议的工作还将为参与该项目和/或其延续项目的研究生和博士后创造培训和就业机会。