"Intelligent monitoring for elderly and the disabled using vision , depth, motion and audio sensors"

加拿大人口正在老龄化。2010年，估计有480万加拿大人年龄在65岁或以上。到2051年，大约四分之一的加拿大人预计将达到65岁或以上。由于加拿大人口老龄化的趋势如此明显，因此需要投入大量资源为老年人提供服务。这一建议的前提是，住在自己家里可以提高老年人的生活质量，增加照顾者的选择，减少对保健资源的压力。该研究计划旨在提供一种低成本的智能家居监控解决方案，以降低无人值守紧急情况的风险。拟议的研究将推进使用视频，深度，运动和音频传感器来确定被监控家庭内居住者的健康状态的最新技术水平。 为了实现这一目标，在短期内，该提案将a）开发用于整合来自多个视频和深度相机的数据的算法，并获得对个人的位置、身份和姿态的准确和一致的估计，B）开发基于物理的模型以实时估计来自视频和深度相机的姿态，以及c）开发可以结合位置的跟踪算法，身份和姿态信息以及数据运动和音频传感器。从长远来看，该提案将研究使用被动红外摄像机来估计生命体征，开发一个动作识别系统，并创建一个全尺寸原型来验证算法。这项研究还将为数据访问制定一个全面的安全模型，并制定隐私指南和审计机制，以确保满足隐私和安全要求。 通过实现短期和长期目标，预计将对活动认可领域做出重大贡献。这些进展，结合对老年人监测的关注，将为一个迫在眉睫的问题提供重要的工程解决方案，这个问题肯定会在未来几十年消耗大量的国家资源。

The Canadian population is ageing. In 2010, an estimated 4.8 million Canadians were 65 years of age or older. By 2051, about one in four Canadians are expected to be of age 65 years or older. With such a significant bias towards an ageing population in Canada, an enormous amount of resources need to be devoted to the services to cater for the elderly. This proposal is based on the premise that living in their own home increases the quality of life of elderly, increases the options for caregivers and reduces the pressure on health-care resources. This program of research aims to provide a lower-cost, smart home monitoring solution that will reduce the risk of an unattended emergency. Proposed research will advance the state of the art in using video, depth, motion and audio sensors to determine the state of well-being of occupants within the monitored home. To achieve this goal, on short-term this proposal will a) develop algorithms for integrating data from multiple video and depth cameras and obtain accurate and consistent estimates of location, identity and pose of an individual, b) develop a physics-based model to estimate pose from video and depth cameras in real-time and c) develop tracking algorithms that can incorporate location, identity and pose information together with data motion and audio sensors. On the long term, this proposal will investigate use of passive infra-red cameras for estimating vital signs, develop an action recognition system and create a full-size prototype for validating the algorithms. This research will also develop a comprehensive security model for data access as well as develop privacy guidelines and auditing mechanisms to ensure that privacy and security requirements are met. By achieving the short and long term objectives, it is expected to make a significant contribution to the area of activity recognition. Such advances, when combined with the focus on elderly monitoring will yield significant engineering solutions to a looming problem that is sure to consume significant national resources in the coming decades.