August 8, 2023 — Southwest Research Institute is launching its new Engine for Automatic Biomechanical Evaluation (ENABLE™) markerless biomechanics system during the American Society of Biomechanics (ABS) Annual Conference, August 8-11, in Knoxville, Tennessee. ABS attendees can see a demonstration by visiting Booth No. 11.
ENABLE is a user-friendly markerless motion capture system that leverages artificial intelligence, computer vision algorithms and biomechanical modeling. The key advantage of ENABLE is it efficiently captures motion without requiring physical body markers attached to a human subject.
“Following years of internal research and applied client projects, SwRI has made ENABLE commercially available for biomechanical analysis in sports, medicine, research and other human performance applications,” said Institute Engineer Dr. Dan Nicolella. “We are excited about all the potential applications from sports to military medicine.”
SwRI Research Engineer Ty Templin will present a paper titled, “Investigation of Pitching Mechanics to Reduce Injury Risk Using Optimal Control,” which discusses using ENABLE and human performance optimization to analyze the performance and injury risk of baseball pitchers.
“With so many open questions about the underlying mechanisms driving pitching injury and actions that can prevent injury, we wanted to design an optimal control framework to investigate the cause-effect relationship between pitching mechanics and injury risk,” Templin said.
ENABLE combines novel artificial intelligence algorithms with OpenSim biomechanical modeling to analyze the properties of human physical motion, or kinematics, from standard video. Sometimes known as the geometry of motion, kinematics enables the mathematical quantification of walking, running and other precise physical movements.
Historically, sports scientists have relied on the tedious process of attaching physical markers to a human subject at precise locations on the body to capture kinematics data with special infrared cameras. ENABLE circumvents physical markers by leveraging novel computer vision algorithms that process standard video. An SwRI-developed neural network identifies and tracks over 80 virtual points on the human body including the locations of wrists, elbows, knees, ankles and other joint positions.
“We trained ENABLE to see those locations in views from multiple cameras. ENABLE robustly computes the 3D location of each virtual point on the body to accurately map the subject’s motion,” said Omar Medjaouri, an SwRI research engineer who specializes in computer vision. “Combining computer vision and biomechanics, ENABLE creates a personalized biomechanical model of each subject and uses this model to constrain the resulting movement to the physical reality of the human body to output kinematics that are highly accurate.”
SwRI recently published a paper titled, “The Effect of Synthetic Training Data on the Performance of a Deep Learning Based Markerless Biomechanics System” detailing how ENABLE, using highly accurate artificial intelligence training data, provides the same level of accuracy as a marker-based system. Visit https://enable.swri.org to download the paper.
“We are excited to demonstrate how synthetic training data can be used to customize this powerful biomechanical tool,” said Nicolella, who co-leads SwRI’s Human Performance Initiative. “We look forward to learning how our data and results can be applied within the biomechanics community to resolve an array of human performance issues.”
ENABLE was developed through SwRI’s Human Performance Initiative, which applies a multidisciplinary scientific and engineering approach to better understand and quantify the complex biomechanical and physiological components of physical performance. SwRI scientists and engineers represent diverse technical backgrounds that include biomechanics, computer science, machine learning, systems engineering, sensor fusion, biomedical engineering, physics, statistics and applied mathematics.
To learn more, watch a video https://youtu.be/49DM6aqxDag or visit https://hpi.swri.org.