I-Corps: Vision analysis system using inferred three-dimensional data to analyze and correct a user’s pose in relation to 3D space

这项I-Corps项目的更广泛影响/商业潜力是发展康复技术，重点是增加以家庭为基础的锻炼方案，以精确恢复活动能力。目前，越来越需要可获得和一致的物理治疗支持，而目前的工具导致依从性差，最终恢复效果差。提出的技术提供了人体的分析，以促进康复的物理治疗患者在诊所和在家里通过音频/视觉反馈和纠正指导。该技术旨在为护理人员提供即时纠正和同步进度，而姿势分析和实时指导系统则在锻炼过程中提供信心。目标是通过改善获得个性化康复的机会来促进更好的健康结果和改善生活质量，从而可能减少医疗保健差距和知识渊博、可获得的护理的成本。这个I-Corps项目的基础是开发一种用于物理康复的软件工具，该工具解决了在物理治疗中为患者独立进行的练习。目前，临床医生受到家庭锻炼工具的限制，这些工具没有可定制的功能。所提出的视觉分析系统使用推断的三维数据来分析和纠正用户在三维空间中的姿势。该技术包括一个机器学习（ML）算法，可以动态推断人体姿势，并根据需要提供纠正措施。此外，所提出的工具利用深度学习方法，通过从结果中学习和确定持续恢复的引人入胜的技术来不断改进。目标是提供高精度的家庭支持，补充身体恢复，直接影响活动和治疗目标。提出的技术提供实时指导、纠正训练和综合进度跟踪，这可能会显著提高家庭锻炼的有效性。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The broader impact/commercial potential of this I-Corps project is the development of rehabilitative technology, focusing on augmenting home-based exercise regimens for precise mobility recovery. Currently, there is a growing need for accessible and consistent physical therapy support while current tools lead to poor adherence and ultimately poor recovery outcomes. The proposed technology provides an analysis of the human body to encourage recovery for physical therapy patients both in-clinic and at home through audio/visual feedback and corrective coaching. The technology is designed to provide instantaneous corrections and synchronized progress with care providers, while the pose analysis and real-time guidance system provides confidence during exercise sessions. The goal is to facilitate better health outcomes and improved quality of life by improving access to personalized rehabilitation, potentially reducing healthcare disparities and cost of knowledgeable, accessible care. This I-Corps project is based on the development of a software tool for physical rehabilitation, that addresses independently performed exercises for patients in physical therapy. Currently, clinicians are limited by home exercise tools that do not have customizable features. The proposed vision analysis system uses inferred three-dimensional data to analyze and correct a user’s pose in relation to 3D space. The technology includes a machine learning (ML) algorithm to dynamically extrapolate human pose insights and offers corrective action as needed. In addition, the proposed tool leverages a deep-learning approach that continues to improve through learning from outcomes and identifying engaging techniques for continued recovery. The goal is to provide high-precision support at-home that complements physical recovery and directly impacts mobility and therapy objectives. The proposed technology provides real-time guidance, corrective coaching, and integrated progress tracking, which may significantly improve the effectiveness of home-based exercises.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.