Artificial Intelligence and The Sports Industry

Artificial intelligence (AI) is changing our lives, and it is everywhere!

The sporting industry is one of the fields that took advantage of the era of AI transformation. Either in sports entertainment or coaching, we can always find the effect of AI everywhere. The combination of sensor technology and AI can help in many sports applications. It can improve the technique and the performance of the players, provide real-time analysis, and deliver personalized training programs for athletes. 

 

In this short blog, we’ll see how advances in AI domains such as Computer Vision, Natural Language Processing, and Time-series Analysis help develop better and more robust sports-related applications.

“AI is the new electricity,” Said Andrew Ng, Associate Professor of AI at Stanford University, at  a conference on artificial intelligence

AI in Sports Entertainment

AI can provide great sporting event analytics via computer vision techniques like object detection and action recognition. For example, AI can detect the ball’s position as well as players in any video game. It can also recognize the actions of each player during the game.


Improvement in deep learning algorithms, especially the convolutional neural network (CNN), made AI-based algorithms possible to detect each player’s position and action at any frame of a sports video. One of the most popular object and movement detection tools is DETECTRON2, released by Facebook. Openpose is also one of the most popular open-source pose estimation technologies that can detect the human body’s key points.

Showing how Object Detection work in sports

Object Detection

Showing how the ai sees the Key Point Detection

Key Point Detection

Further, more tasks have room to be automated. Machines can now generate natural language describing what happened in a sporting event in the form of posts, tweets, or text messages that humans can understand. GPT-3 and GPT-2 are examples of models that have been trained on a massive amount of text to generate such tasks. 
Example of a GPT-2 generated text given a sentence.

Example of a GPT-2 generated text given a sentence.

We can even leverage that advance in the form of chatbots. Chatbots can easily help fans with their queries, such as attending events, reserving tickets, parking, and seatings. Chatbots were used before by Tampa Bay Lightning (an Ice hockey team) and Sacramento Kings (a Basketball team). The power of chatbots is that they can handle thousands of users concurrently without needing humans, and they can work on numerous platforms like Facebook, WhatsApp, Discord, etc. 


Did you also know that AI plays a significant role in advertisements? IBM Watson has been used in Wimbledon 2017 to recommend the optimum time slots to display ads to earn the optimum engagement given the player’s actions, spectators’ emotions, and commentator’s language.

AI in Coaching

Artificial intelligence in sports is an excellent tool for coaches as well as individuals. On a professional level, coaches can improve players’ performance by looking into players’ statistics, and observing and creating metrics, such as speed. By using AI and Machine Learning in IoT-powered sensors, coaches can make better decisions for their athletes and minimize the risks of injuries. Such sensors and devices connected to specific apps can then track the athletes’ health and performance and collect athletic data for analysis.

Apps for tracking activity and wearable technology are the most common types of technology used by athletes and sport and fitness enthusiasts, with popularity among younger adults, particularly younger women. However, the future of wearables is going beyond just tracking. By applying Machine learning and AI, one can learn from performance data and derive in-depth recommendations to provide tailored programs for athletes and fitness enthusiasts that increase their performance and workout efficiency to reach their set goals.

Inertial measurement unit (IMU)

IMU, a collection of sensors, including an accelerometer and gyroscope, is available in many of today’s belts and watches, facilitating user data recording, which powers AI coaching.

Sample of Sensors

Racefox is an example of an AI-based running and skiing coach. As a mobile app, it provides real-time voice coaching to guide you with your personalized training plan. When you first use Racefox, it asks you to do a 15-minute exercise.

This exercise will provide the app with the needed data to understand your fitness level and heart rate zones. The AI coach will generate many KPIs such as step frequency, ground contact time, drag, impact, and pounce based on the sensor data such as heart rate and accelerometer data. By analyzing the data, the AI digital coach will then coach you to improve your performance, avoid injuries, and reach your goals.

 

Another example is Qualysis which is one of the leading motion capturing companies that help in AI coaching. It provides analytics and statistics about user movements based on high-end optical motion engines alongside an advanced analysis backend. 

how ai works in sports

Qualisys

Application with AI in sports industry

Racefox

AI coaching provides ample alternatives to human coaches and provides personalized training plans for everyone.

Everybody Makes Mistakes

Even AI and ML make mistakes, but they learn from these mistakes and become smarter so they do not get repeated

Although AI has enhanced human experiences in significant ways, it might humorously fail in some tasks. Here is an example of how an AI-driven camera that moves based on ball tracking has misclassified the referee’s shiny bald head as the ball and kept focusing on it, leaving the real ball and players on the other side of the field.

In the future, you may not even need a coach to communicate the AI decision as the machine will communicate it directly to the players using realistic “human” communication and added emotion. This technology may not be too far away in the distant future and is currently being experimented with within the NFL. Using deep learning, sporting codes can uncover strategic insights that may not have been previously achievable and will differentiate teams willing to invest.

Optomatica and Smart Sports

Revolutionizing the health, sports, and fitness world with our broad expertise in Artificial intelligence, Optomatica provides smart software solutions, such as applications providing digital coaching and guidance using AI techniques. Whether you are interested in developing a running, skiing, cycling, fitness training, or rehabilitation app, we can partner with you to develop a solution that helps trainers stay stronger, healthier, and injury-free.

Author: Aliaa Rassem & Mohamed Sayed
Editor: Sherif Adel