Game Vision
RESEARCH

VISÃO DE JOGO

An idea born during the 2014 World Cup — and that technology has only now caught up with.

In 2014, while editing the World Cup for Fox Sports, Ricardo spotted something no broadcast was doing—and still isn't today. The idea stayed shelved for over a decade, waiting for the technology to catch up.

It has matured. With segmentation models like SAM 3 and detection networks trained specifically for soccer — players, ball, field, movement — it's now possible to understand a match in real time, frame by frame, with precision that was science fiction in 2014.

What exactly are we going to do with this? It's still a secret. But someone who spent 28 years inside sports broadcasts knows where it hurts — and where no one has looked yet.

Computer vision

SAM 3 + soccer-specific models

Real time

Players, ball, and field understood frame by frame

28 years of broadcasting

The perspective of someone who edited the World Cup and Olympics from the inside

Originality

Nothing like it exists — in 12 years, no one has done it

EM IMAGES

O palco: 90 minutos de dados em movimento
O palco: 90 minutos de dados em movimento
The ball as the machine sees it
The ball as the machine sees it
Every play, thousands of tracked points
Every play, thousands of tracked points
Da prancheta ao algoritmo
Da prancheta ao algoritmo
New angles, new questions
New angles, new questions
The room where the broadcasts are born
The room where the broadcasts are born
Where it all begins: the street
Where it all begins: the street
The moment the analysis is trying to explain
The moment the analysis is trying to explain
Physics, rain, and a fraction of a second
Physics, rain, and a fraction of a second

WANT TO FOLLOW ALONG CLOSELY?

Create your free account and be notified of every update to this project — before everyone else.