On Monday, the iMerit team will be arriving in Santa Clara, California to join the many innovators, technologists, and engineers from the computer vision industry at this year’s Embedded Vision Summit. The Embedded Vision Summit brings product and application developers, business leaders, investors, and entrepreneurs who are all focused on embedded vision together to see the latest in practical technology. Attendees get to see the latest in practical technology and […]
On Monday, the iMerit team will be arriving in Santa Clara, California to join the many innovators, technologists, and engineers from the computer vision industry at this year’s Embedded Vision Summit.
The Embedded Vision Summit brings product and application developers, business leaders, investors, and entrepreneurs who are all focused on embedded vision together to see the latest in practical technology. Attendees get to see the latest in practical technology and dataset services to bring visual intelligence into cloud applications, embedded systems, mobile apps, wearables, and PCs.
Along with showcases of the latest computer vision products and services, attendees will get to hear from industry experts in five in-depths tracks on technical insights, business, enabling technologies, and fundamentals of visual intelligence. Steven Cadavid, President and Founder of Kinatrax, will be presenting on their markerless motion capture system that computes kinematic data of an in-game baseball pitch.
Markerless Motion Capture System Captures Kinematic Data of MLB Pitchers
KinaTrax’s Markerless Motion Capture System consists of a variety of imaging devices that are mounted throughout the baseball stadium to capture the detailed movement of pitchers. They develop 3D kinematic models that are used by teams to monitor pitchers’ performance and to prevent injury. Currently, these devices are mounted at the home stadiums of the Cubs and the Tampa Bay Rays, along with another undisclosed stadium.
Computer Vision and Machine Learning algorithms are used to capture the biomechanics of a pitcher at over 300 frames per second. The video is recovered in 3D and reconstructed frame by frame, producing an image for every motion within the pitch sequence which are then annotated at 20 joint centers. KinaTrax leverages iMerit’s team of computer vision data experts to provide on-demand and scalable annotation resources from end-to-end through KinaTrax’s data analysis workflow.
To create the 3D models, “iMerit teams process in-game footage and prepare it for analysis. We then annotate images and videos to create 3D models of the players. iMerit integrates directly with KinaTrax systems, facilitating a seamless handoff of data. In the past year, iMerit has built a precise 3D model for approximately 300 pitchers across all 30 MLB teams, covering most pitching personnel,” said Jai Natarajan, VP of Technology at iMerit.
Computer Vision Algorithms and Humans Work Together to Create Baseball’s New ‘Moneyball’
The idea of leveraging in-game data to enhance pitcher performance “will make a profound difference in major league baseball, it will change the game,” said Cadavid. Billy Beane revolutionized baseball with his analytical, evidence-based approach to selecting players dubbed ‘Moneyball’. KinaTrax’s offering is supplying team management with the data they need to make the best decisions about a pitcher’s health.
The 3D Kinematic models can be used to generate comprehensive and customizable biomechanic reports. According to Steven Cadavid, “The primary uses for it are evaluating the mechanics over time, the performance enhancement and injury prevention component. From an injury prevention, the datasets we’re collecting are really unprecedented.”
KinaTrax is taking a different approach to motion capture technology by leveraging video annotation to gain the data required to build 3D kinematic models. This means that the subjects, in this case, the pitchers, don’t need to wear markers to capture the data. This key element to KinaTrax’s technology allows them to not only capture training data but in-game data as well. Combining this with iMerit’s on-demand dataset service offering, humans in combination with technology are revolutionizing the game.
Want to Learn More about Data-Driven Baseball?
If you are attending the Embedded Vision Summit May 1-3, be sure to check out Cadavid’s talk on Using Markerless Motion Capture to Win Baseball Games. Also be sure to come by the iMerit booth to grab some Philz coffee. If you are unable to attend, check out this video.
Want to learn more about how iMerit can help your Computer Vision technology? Get in touch with our team today.