In an effort to enhance offside violation detection in the Premier League, Genius Sports has partnered with the UK's top soccer league. Utilizing cutting-edge technology and machine learning models, this collaboration aims to provide more accurate and efficient offside calls during games.

Offside violations can be challenging to determine, especially in situations where players are closely grouped or obstructed from view. Traditional video assistant referee (VAR) systems have been employed to address these complexities by leveraging cameras and machine learning algorithms. However, these systems have faced limitations, often resulting in extensive delays and human errors.

Recognizing these challenges, FIFA, the global governing body for soccer, introduced limb-tracking technology powered by machine learning and embedded soccer ball sensors in 2022. This software was capable of tracking 29 points on players' bodies; however, concerns persisted regarding precision and potential errors in in-game calls.

Genius Sports brings its innovative technology known as "Semi-Assisted Offside Technology" (SAOT) into play. This technology forms part of GeniusIQ, Genius Sports' comprehensive system that not only supports game officials but also provides engaging fan experiences through real-time dynamic graphics and visualizations.

SAOT offers enhanced 3D player renders, enabling officials to precisely define the offside line and accurately determine each player's position in relation to it. To achieve this level of detail and accuracy, Genius Sports employs an extensive network of cameras.

Genius Sports initially considered transitioning from high-resolution 4K cameras to more cost-effective solutions. After exploring various options, they found that iPhones provided the desired balance between affordability and performance. Familiarity with the iPhone platform among Genius Sports' employees, particularly in software development, played a significant role in this decision.

To ensure comprehensive coverage of both the pitch and sidelines, Genius Sports deploys approximately 24 to 28 iPhone devices, often arranged in custom rigs holding two phones at a time. These rigs are strategically angled to provide optimal visibility and coverage.

This extensive deployment results in an impressive amount of data points - between 7,000 and 10,000 - enabling the creation of a detailed 3D virtual mesh for each player. This high volume of data points allows for tolerance in dealing with potential issues such as lighting conditions or other environmental factors that may impact camera visibility.

In addition to their high frame rate recording capabilities - Genius Sports records at 100fps, with potential testing at 200fps - iPhones also offer local computer vision processing. All captured data is transmitted to an on-premise server for processing through GeniusIQ's advanced computer vision and predictive algorithms.

These algorithms are designed to identify individual body parts, even down to players' fingers, and predict their positions when they are obstructed from view. Genius Sports trained its system using several seasons of soccer matches to ensure accurate identification and tracking.

Ultimately, this technology aims to determine the precise positioning of every player in relation to each other, the ball, and the goalkeeper. Offside calls are made when the ball leaves the offensive player's foot; therefore, capturing more frames increases the likelihood of accurately identifying the exact moment of an infringement.

While the official rules of soccer provide specific guidelines for determining an offside violation, it remains unclear whether GeniusIQ's more granular approach offers superior performance compared to existing VAR systems. Genius Sports refrained from providing specific performance comparisons but emphasized that other VAR systems may only track 30 or 40 points of the body or rely on "center of mass" tracking, resulting in rough stick figure representations of players.

As Genius Sports' offside detection technology prepares for full implementation in the Premier League before the end of this year, fans and officials alike eagerly await its potential impact on the accuracy and efficiency of offside calls during games.