To help aspiring baseball players analyze and improve their technique, NVIDIA Inception member wrnch, in collaboration with ProPlayAI, developed a GPU-powered biomechanics app for pitchers.
The app is designed to help people stay fit, improve performance, and coach players on the technical details.
“Our approach allows athletes to perfect their pitching mechanics, even when they cannot access facilities, other players, or coaches due to social distancing,” said ProPlayAI co-founder and MLB pitcher, Nate Pearson. “It democratizes coaching and makes biomechanics as part of your daily training, without the need for specialized cameras or sensors at a lab.”
To run the GPU-powered models, players set up a standard smartphone camera to capture their moves. Once recorded, the videos are uploaded to ProPlayAI servers where the data is analyzed by wrnch’s AI computer vision engine running on NVIDIA GPUs.
The processed data is then sent back to ProPlayAI which uses additional proprietary algorithms to analyze posture and provide users a report of full-body 3D kinematic data, including arm speed, deceptiveness, and efficiency.
“ProPlayAI is a wonderful example of wrnch’s mission to enable companies to easily add human-centric computer vision to their applications,” said Dr. Paul Kruszewski, founder and CEO of wrnch.
The wrnch AI engine uses advanced neural networks trained on large, highly diverse data, to analyze videos and perform human pose estimation. Training and inference on both models are powered by NVIDIA GPUs.
NVIDIA Inception is a program that nurtures AI and data science startups with go-to-market support, expertise, and technology.