Markerless Biomechanics to Investigate Performance
Abstract
Over the course of the program the base technology for an automated medical training feedback tool was created. This included a multi-stage hand tracking system which can track detailed subject hand kinematics from video without the need for any sensors or markers to be attached to the subject. A graphical user interface was created to incorporate this tracking pipeline along with the necessary tools for camera calibration as well as packaging for use by nontechnical users. Several different approaches to automatically generate feedback were developed, and a test data set of medical students, residents, and attendings performing suturing was collected. Resulting algorithms were able to successfully identify suturing errors within the suturing data set automatically.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 30, 2022
- Accession Number
- AD1219294
Entities
People
- Daniel Nicolella
- Travis Eliason
Organizations
- Southwest Research Institute