NFL, AWS unveil winning entry in Contact Detection Challenge

The NFL and Amazon Web Services reported that the winning entry in its Contact Detection Challenge, a computer vision and machine learning contest to aid in the prediction of injuries, marked a 31% improvement in automatically identifying on-field contact. Nghia Van Ngoc Nguyen, a Vietnamese data scientist now based in South Korea, submitted the winning entry that included a new mechanism for detecting contact when players make contact with the ground.

Google-owned Kaggle hosted the crowdsourced competition in which users could access AWS Next Gen Stats and multiple video angles. Nguyen is a regular high finisher in Kaggle contests who earned $50,000 as top prize. Engineers at AI cloud company H20.ai comprised the runner-up entry, winning $25,000, while another $25,000 was distributed to third, fourth and fifth places.

This was the third annual public data competition held by the NFL and AWS as part of their Digital Athlete collaboration. A record 939 teams from 71 countries submitted entries.