Computer Vision App


UCReport, 3STEP Sports


2 years


American Football

key technologies

Computer Vision, Generative AI

project overview

The UCReport, the premier underclassman scouting database for American Football, asked us to develop a computer vision platform to calculate the speed of football players from game film. We built a comprehensive computer vision system to do just that, allowing UCReport analysts and coaches at 68 of the 69 Power 5 universities to access the data for recruiting.

project videos

project images

App - Player Tracking
App - Speed Metrics

detailed description

This project was broken into two parts over the course of a two-year engagement. The first part was the MVP launch, which was completed in the spring of 2022. The second part was the V1 launch, which was completed in the fall of 2023.

This is a traditional process we follow with our clients: (1) build a minimum viable product (MVP) to get the product to market as quickly as possible, (2) solicit sales, feedback, and traction with the MVP, and (3) iterate on that product to add features and improve the user experience.

Sometimes an MVP won't move beyond that stage, but in this case, the MVP was successful and the V1 launch was completed. The V1 launch included a number of improvements, including a more accurate speed metric, a more user-friendly interface, and a more robust backend.

The Twelve55 team used a combination of computer vision and machine learning tools to complete this project, with the most important product decision being the balance of machine and human input to ensure the accuracy of the speed metrics. The result was a platform that is accurate, scalable, and user-friendly.