“Some people worry that artificial intelligence will make us feel inferior, but then, anybody in his right mind should have an inferiority complex every time he looks at a flower.” – Alan Kay

Welcome to the Tech issue. Technology is the heartbeat of our business and we are working tirelessly behind the scenes to make sure our content offering is up to our standards. We had a productive and insightful visit to the Pixellot HQ in Tel Aviv, as we continue to find ways to improve our tech. We also celebrated crossing 160 000 app downloads, a significant milestone only halfway through the year.

In this month’s newsletter, we want to focus on giving you a better understanding of the underlying Artificial Intelligence (AI) technology used in the production of games from our Pixellot cameras. For me, this is arguably the most exciting initiative of our business. This area has also been where the most of our criticism has come from, and understandably so.

Despite the criticism, below are a few strategic reasons for rolling out the AI tec:

• Scalability – To roll out the project to the number of schools we want to reach, we need to find solutions that will not need a lot of human intervention to produce games.
• Hardware improvements – the zooming capability of the lenses in the camera will improve over time, resulting in improved-quality productions
• AI Software upgrades – When KPI’s are identified where the AI gets the production wrong consistently, programming code can be written to identify these patterns and then be forced to produce that pattern of play correctly.
• Full data set of a match – The four lenses of the camera is calibrated to capture the content from the full court or field. This raw data is then used by the AI software to track the ball and produce the game. When a new version of the AI software is released the Video on Demand (VOD) production of the game will look completely different to the previous software version’s VOD production.
• PANO – having the full data set of a match available, you have the option to watch the VOD game either in the traditional HD mode (produced by the AI software) or in a panoramic mode (PANO) where you have access to the raw data and can determine where you want to focus on during the game. You don’t have to follow the ball and become the producer. Coaches can follow specific players which might not be on screen if you watch it in HD mode.
• Future data analysis – Currently the AI software can trace the ball. It will be very soon that players can be tagged enabling you to track a specific player for the full game. Speed and distances run by players or teams collectively, might be available in future. Not only of future games, also of the games where the raw data has already been captured.

I hope that you will have patience with us on this journey. We know that the AI software can miss important moments that you care about most.  We hope that you understand that we can only provide you with this content using this affordable and scalable solution, understanding that it has future benefits that we might not even think about today. What an exciting world we live in!

– Gerhard Steyn

Pixellot has partnered with ITV Africa and Nashua to build and support the SuperSport Schools project. Pixellot’s products have a low bandwidth requirement which provides great results in low coverage areas which are prevalent in South Africa.

What makes this solution unique and valuable to the SuperSport Schools project, is the ability to cover multiple sport types without the need for human intervention. This is made possible by each sport type having its own unique AI algorithms. Each algorithm is uniquely created and tailored to suit the different sport types, and their various game play styles.

The AI is ever-evolving, and we are seeing constant improvements month by month in the Pixellot system’s ability to track, play by play, the different sport types we see here in South Africa. Our latest opportunity for development came in relation to rugby. The nature of the game required exclusive updates to the AI technology to accurately track the dynamic style of rugby. Rugby is a difficult sport for the Pixellot systems to learn, as the ball is often not visible in rucks, mauls, and scrums. However, with improvements to the AI and software of the rugby systems which came in mid-June, we will see a vast improvement in the system’s ability to track the game of rugby more accurately.

There are ever increasing opportunities for Pixellot systems to evolve and meet the challenges of the future.

“AI is neither good nor evil. It’s a tool. It’s a technology for us to use.”

– Oren Etzioni