AI Detection Software has become increasingly advanced, and with advancement comes opportunity.
I have been working with different computer vision detection models for the last 6 years as part of my research in computational biology and cancer AI. As such, I have had the opportunity to use and build some of the state the art detection and segmentation software for various diseases both in pathology and radiology.
I wanted to take this project as an opportunity to apply some of the learnings and algorithms I had built up over the years. Innovation is often transferred across domains, and I saw a chance to create value in the world of sports analytics.
Creating player detection on the court is no easy task, especially when you have to locate the players, identify team belong, all with a single feed moving camera. The project I present is a culmination of 7 different AI models, all working together to provide a cohesive data engine for player tracking with the bare minimum equipment.
I spent hours labelling thousands of data points to track players, balls, and even understand how the court translates to the 3D world. The result is a rich source of data that is accessible to anyone that can watch a NBA feed on their own computer, opening the door to advancement in coaching, analytics, sports betting, broadcasting, and even personal entertainment.
Log in or sign up for Devpost to join the conversation.