AI-Powered Real-Time Betting: Revolutionizing the Casino Experience

Customer Overview

Our client, a leading online platform, specializes in providing personalized betting experiences to sports enthusiasts. It offers a diverse range of sports events and attracts a significant user base globally. The client wanted to take the casino experience to the next level with a progressive idea.

Project Overview

Our client wanted to build a fantasy league and innovative digital betting platform for live poker games. They wanted to provide 2nd level of revenue to casinos by facilitating betting on the existing games of betting. They intended to revolutionize the casino experience for global users by allowing them to bid on poker players during live games.

Challenges

Building a platform that provides accurate real-time poker game data for betting decisions, handles high volumes from multiple casinos, and complies with regulations.

• The platform must display vital information, like players, cards, pot, current hands, etc., about live games in a user-friendly way. Also, the display of this information should be in real time.
• As betting involves money and the platform users will base their decisions on the displayed vital information like cards, players, etc., it is essential to ensure data accuracy.
• The platform must support integration with different video streaming systems of casinos and consume live video streams of poker games.
• The platform must ensure regulatory compliance and responsible gambling practices.

Solution

We developed a sophisticated web application that while streaming live games from multiple casinos at scale, provides their vital tables and leaderboard. It allows users to place bets during live games by providing accurate real-time data.

The solution focused on the following key areas:

• To stream live games from multiple casinos at scale, we linked the AWS Kinesis Video Stream channel to each casino’s poker game. The system we build configures and adds multiple casinos and poker games going on concurrently within each casino.
• AWS Sagemaker processes Real-time AWS Kinesis Video Streams and carries out model inferencing for Object Detections and OCR. It processes every 10th frame of the video for reliability of information.
• To capture accurate information in real-time from the video of live games, we have trained Deep Learning YOLO-based vision models to detect object sections like players, pots, cards, etc. Additionally, we have utilized a Deep Learning-based custom-built OCR model to grab textual content from the detected objects in real-time.
• We achieved horizontal scaling of model inferencing processes by deploying both the Object Detection and OCR models using AWS Sagemaker.
• After real-time processing of the video frame, we republish the processed stream over HLS with timed metadata embedded in a JSON file (representing the game’s current state) along with the video frame. It keeps the video stream and the game’s current state in sync.
• We utilized Angular framework to build the web application and hosted it on AWS Cloudfront.

Benefits

• The developed solution provides 2nd level of revenue to the casinos, thus increasing their growth and profitability.
• The technology that we have created, can be applied to any of the live sports and a fantasy league platform could be created.
• Personalized recommendations, interactive features, and gamification elements increased the platform’s user engagement.
• Our client solidified its industry leader position and established a reputation for innovation, reliability, and compliance.

Technology and Integration

• AWS Kinesis Video Stream
• AWS Sagemaker
• YOLO
• PaddleOCR
• OpenCV
• Python
• Angular JS

Industry

• Online Gambling Industry