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

The intellectual merits of the proposal include (a) a single-view motion capture framework that combines pose detection and pose refinement (b) various enhancement and risk mitigation plans to meet the stringent need as a medical diagnostic tool. These include the construction of anatomically-correct six degree of freedom lower extremity joints, robust sensor fusion algorithms, and system development to allow user interactions; and (c) a thorough quantitative evaluation of our approach’s accuracy for medical applications, using commercial markerbased motion capture system to generate the ground truth. To further increase success of the project, we further limit the assessment to the lower extremity, which reduces the complexity of motion estimation, but nevertheless maintains its usefulness in many medical conditions. The broader impact of this proposal will be a transformational improvement in the technology that is available to physicians and physical therapists to diagnosis, treat, and monitor patients with a movement disorder. These disorders encompass a wide range of areas from children with cerebral palsy, to athletes at risk for serious knee injuries to the geriatric patient with neurologic impairments such as Parkinson’s disease or a stroke. The current state of the art for technology to measure 3D motion is expensive, cumbersome, and not widely available to the medical professional. The development of a portable, low cost, single camera system will give medical professionals accurate and precise 3D measurements of movement dysfunction in any environment. This would have important clinical ramifications in the ability to diagnosis, treat, and prevent injuries. The medical professional would also be able to monitor the individual’s progress, both in the clinic and remotely from their home, decreasing the need for costly office visits. In addition the high-level 3D metric body configuration from our system could also be used for a whole range of health and security monitoring applications, without the usual privacy concerns associated with video imagery.