MirrorMatch: Real-Time Detection of Repetitive Movements Using Smartphone Camera
Description/Abstract/Artist Statement
Keeping track of one's personal fitness is challenging and a topic of ongoing research. For years, sports enthusiasts have relied on themselves, trained professionals, and systems implemented with wearable technology to assess the correctness of their movements. Unfortunately, these error-prone and costly practices can lead to misinterpreted progress or an injury related to the athlete’s workout routine. An automated approach to supervising exercise movements could help fitness enthusiasts and athletes log their physical activities and prevent injuries in a cost-effective manner. We propose MirrorMatch, a motion tracking system designed to assist athletes in monitoring their physical fitness. Our approach provides real-time movement analysis and feedback through the use of commercial off-the-shelf smartphone cameras, machine learning techniques, and image processing libraries. After the user completes an exercise, they are presented with a detailed workout report. Over time, multiple workout summaries can be combined to generate a progress report, giving the user insight into their performance to account for all development and help prevent any injuries. Unlike existing solutions that use sensors or immobile camera systems, MirrorMatch offers a practical and scalable solution to make fine-grained movement tracking more accessible. Such a system could impact the realm of physical therapy and help facilitate rehabilitation for injured patients by providing the necessary feedback over the course of their treatment.
Faculty Advisor/Mentor
Shubham Jain
Presentation Type
Poster
Disciplines
Body Regions | Graphics and Human Computer Interfaces | Software Engineering | Statistics and Probability | Systems and Integrative Physiology
Session Title
Poster Session
Location
Learning Commons, Atrium
Start Date
2-8-2020 8:00 AM
End Date
2-8-2020 12:30 PM
MirrorMatch: Real-Time Detection of Repetitive Movements Using Smartphone Camera
Learning Commons, Atrium
Keeping track of one's personal fitness is challenging and a topic of ongoing research. For years, sports enthusiasts have relied on themselves, trained professionals, and systems implemented with wearable technology to assess the correctness of their movements. Unfortunately, these error-prone and costly practices can lead to misinterpreted progress or an injury related to the athlete’s workout routine. An automated approach to supervising exercise movements could help fitness enthusiasts and athletes log their physical activities and prevent injuries in a cost-effective manner. We propose MirrorMatch, a motion tracking system designed to assist athletes in monitoring their physical fitness. Our approach provides real-time movement analysis and feedback through the use of commercial off-the-shelf smartphone cameras, machine learning techniques, and image processing libraries. After the user completes an exercise, they are presented with a detailed workout report. Over time, multiple workout summaries can be combined to generate a progress report, giving the user insight into their performance to account for all development and help prevent any injuries. Unlike existing solutions that use sensors or immobile camera systems, MirrorMatch offers a practical and scalable solution to make fine-grained movement tracking more accessible. Such a system could impact the realm of physical therapy and help facilitate rehabilitation for injured patients by providing the necessary feedback over the course of their treatment.