The new platform developed by our team was found to be 97.6% accurate – Rohit Handa, PGP AIML

MBA graduate, PG in AI/ML and Computer Science Engineer with 20+ years of IT experience, Rohit Handa is an alumnus of the PGP Artificial Intelligence and Machine Learning Course offered by Great Learning. Read further to learn about his journey with Great Learning.

• Vast experience in Business & Technology Consulting, Solution Development & Delivery for global clients in Media & Entertainment industry, including transformation, development, support, platform migration & legacy modernization programs.
• Certified Data Science & ML, Azure Cloud practitioner
• Driving consulting assignments to implement AI/ML, Azure ML, CNN, and NLP algorithms, customized python/PyTorch programs for large media conglomerates
• Expertise in implementing automation tools like Blueprism, Automation Anywhere, enabling business process automation and workflow management

This is a short yet highly successful story of one of our esteemed US-based clients, a pioneer in internet streaming media and broadcasts delivery software and services. Recently the client contacted our consulting team challenging us with an interesting problem faced by their end client: a renowned UK-based hotel cum casino chain. 

Despite casino’s several great initiatives and strong society’s adherence, their gaming venues were struggling to identify barred people to enter casinos in a real environment as manually monitoring for each person of interest was an arduous and inefficient task — which is very difficult without the right platform and surveillance technology.

The casino was looking forward to helping them develop a highly accurate facial recognition & security alert platform, well-integrated to the UK barred people facial image database as well as casino’s surveillance IP cameras & central monitoring system. 

This new platform was expected to provision the ability to quickly identify barred patrons and automatically notify matches to surveillance staff. As a result, gaming venues’ authorized security teams would have a better chance to enforce gambling bans in gaming venues.

Hence the new platform was anticipated to ensure compliance with responsible gaming regulations, identifying barred patrons, and alerting staff.

A four-step deep neural network-based solution approach was taken to address this unique problem –

1. Pull Barred People from UK State DB and Create a Unique Face Signature– Our platform was integrated into the UK state-provided barred people database, and the integration was set to automatically sync DBs. The face images were processed, filtered, and converted into a unique face signature comprised of mathematical vectors in Euclidean space derived from over a thousand facial features.
2. Connect to IP Cameras — Our platform was connected to camera streams by RTSP (the real-time streaming protocol used for media stream delivery) through the client’s local network.
3. Detect and Recognize faces in live video streams — Our platform was designed using deep neural network techniques to detect and track multiple faces in each live video stream. Each face image was clipped, optimized, and sent for processing. The detected face was converted into a unique face signature. The resulting signature was matched against the UK state database, which can compare millions of signatures in near real-time.
4. Action on Match — If a match is found, the individual is recognized, and the program sends an e-mail and SMS notification to venue staff in less than 60 seconds.

This platform, in partnership with the client’s high-end IP cameras, operating systems & recognized UK state barred face images database, proved to be highly successful in automatically detecting barred patrons inside the gaming venue. 

The new platform, developed by our team, was found to be accurate 97.6% of the time in the barred people recognition by automatic detections, showing the benefits of face recognition technology. The hotel surveillance staff was alerted through emails/texts in less than a min for any barred entries to gaming venues. This could help hotel & casino chains reduce around 25% of the operational expenses spent over penalties incurred due to regulatory violations.

Our platform success was well acknowledged through one of the testimonials from the hotel operations manager.