Developing Computer Vision and Artificial Intelligence for best-quality and cost-effective live sports production and distribution

Project Information

OZeye

Grant agreement ID: 881323

DOI

10.3030/881323

Project closed

EC signature date 4 October 2019

Start date 1 October 2019

End date 30 September 2021

Funded under

• INDUSTRIAL LEADERSHIP - Innovation In SMEs

Total cost

€ 3 419 165,00

EU contribution

€ 2 393 415,50

Coordinated by

OZ EHF
Iceland

Periodic Reporting for period 2 - OZeye (Developing Computer Vision and Artificial Intelligence for best-quality and cost-effective live sports production and distribution)

Reporting period: 2020-10-01 to 2021-09-30

Summary of the context and overall objectives of the project

At OZ Sports we understand the need for sports to enhance the experience for its fans. Our mission is to bring immersive fan experiences to sports fans across the globe. We push broadcast and fan engagement into the new digital age, enhancing fan passion and energy.
Leagues and sports events are missing out on exposure and revenue due to limited broadcasting and commercial opportunities outside of in-stadia sponsorship. The perceived costs of doing multi-camera production creates a barrier to entry. Many leagues and teams across the globe have limited resources to operate and produce matches.
We are changing the economics of sports production by making our latest technologies available to sports federations, leagues and teams at a fraction of the cost. This empowers our customers to deliver engaging and interactive experiences to sports fans around the world. OZ Production delivers high-quality broadcasts for any league, any level and any location.
This carries both importance in the technology space, where we are pushing the boundaries of what is possible and achievable, but also important for sports bodies and the sports fan. The way that fans want to consume content, even niche sports, is critical. All sports need an outlet to reach their audience.
The business focus and the challenges caused by the global COVID-19 pandemic are addressed within this project. The pandemic has accelerated the need for high quality, affordable technology for sports production and distribution and also the need for AI technologies to reduce the number of crew personnel required to produce matches at stadiums. Whilst the sports industry, along with many other sectors, has been struggling, there is an opportunity for agile sports technology businesses like OZ Sports to deliver innovations to the market that provide opportunities for sports leagues, clubs and federations to get back on their feet and start to future proof their businesses.
Within this project we are working towards solving and addressing several key objectives across both AI, content delivery and fan engagement.
● Improving the game experience by using AI to recognize different plays and sequences.
● Optimising hardware to produce exceptionally high quality to broadcasts for all market players
● Detect objects (player, ball etc.) and key incidents and plays in a football match to aid match production and quality.
● Retraining the models to cover sports outside of football.
● Using CDN hosting and delivery to boost performance, scaling and quality of matches to reach large scale audiences across the globe.
● Increase the degree of fan immersion and virtual experiences.
● Enhance replays with better player trajectory.
● Augment the viewer experience with highlights, crowd augmentation, multi-angle images and strategic content overlays.
We are working towards the overall goals of the project and aiming to reach all of the objectives set in a real world situation. We have successfully made our first installation of the OZ system in a football league, which gives us the actual match footage and platform to build upon and demonstrate not only our technical developments but also to aid our commercial and marketing objectives. The project is enabling us to push the boundaries and innovate at an exceptional pace.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

The work performed on the front of AI, and Control Systems focused on dataset development and algorithmic design, developing both control algorithms and image processing algorithms. Following the active learning paradigm, we’ve developed semi-automatic labeling methods, which increased the labeling speed significantly compared to only manual labeling. A prototype of a fully working pipeline has been developed, i.e. from perception to control and production, which means that every new development is able to have a direct impact on the production quality. With this new immersive technical developments we have been able to give live to smaller leagues. Gaining traction from bigger leagues and are now merging in a cooperation with the rightsholders of the Icelandic top league in football. Making it our first league in Europe.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

In terms of both producing and distributing content, we believe we have something one of a kind. We’re leveraging, adapting, and extending state-of-the-art algorithms to automatically produce content. Our intermediate results indicate that we are able to automatically produce content that, most importantly, is pleasant to watch. Following our current path, we expect to increase the automation capabilities and decrease the latency of the pipeline, resulting in higher-quality content with less dependence on manual input.

Detail of one of OZ cameras

View from the control room

An example of a camera arrangement