I-Corps: Assessment of the feasibility of commercialization of a 3D markerless motion capture system with application to physical therapy

低成本和准确的运动捕捉系统是对医生、理疗师和其他医学专业人员诊断、治疗和监测患有下肢运动障碍的患者的技术的根本改进。目前，运动分析系统的成本为数万美元；需要专门的空间和专门的操作培训。因此，执行运动分析的成本高达1000美元/次，而且只有在大型医院才能获得。低成本深度相机的问世带来了以当前成本的一小部分执行运动分析的可能性。在这项提案中，该团队开发了一种临床上可行的3D无标记运动捕获系统，该系统满足了医学界对高准确度和精确度以及便利性和实惠的强烈需求。该团队建议使用商用的深度和彩色相机来准确和精确地定义受伤人群的运动机制。该团队开发了一种有效的人体姿势跟踪系统，实现了相对于基于标记的系统的相对精确度。此外，该团队还开发了在一般场景中跟踪复杂人体姿势的新颖算法，获得了最先进的结果。所提出的系统将建立在现有的新算法流水线上，以实现临床上常用的相对较好的受控运动的更高精度和稳健性。

A low cost and accurate motion capture system is a fundamental improvement in the technology that is available to physician, physical therapists, and other medical professionals to diagnosis, treat, and monitor patients with movement disorders in the lower-extremity. Currently motion analysis systems cost tens of thousands dollars; require dedicated space and special training for operations. Therefore the cost to perform motion analysis is up to one thousand dollars per session and only available in large hospitals. The availability of low-cost depth cameras brings the possibility to perform motion analysis with a fraction of the current cost. In this proposal, the team has developed a clinically viable 3D markerless motion capture system that meets the strident needs of the medical community for a high accuracy and precision as well as convenience and affordability.The team proposes to use a commercially available depth and color camera to accurately and precisely define movement mechanics of injured populations. The team has developed an effective system for human pose tracking that achieves comparative accuracy to marker-based systems. Moreover, the team has developed novel algorithms for complex human pose tracking in general scenes that achieves state-of-the-art results. The proposed system will build upon the current pipeline of novel algorithms to achieve higher accuracy and robustness for the relatively better controlled motions commonly used in clinics.