Facial Recognition Technology Advances

In its simplest form, security cameras record video (facial detection), and facial recognition systems use that video to look for an image match (faces). Current uses also include linking image detection to the ClubsNSW Multi-Venue Self-Exclusion (MVSE) database. 

As early as the 1970s, facial recognition systems started their “image matching” by mapping biometric facial features, like the chin, and then calculated the distance ratio between facial features to create a match. 

These systems quickly evolved to make identifications from less-than-perfect face views and are now good enough to match through impediments such as moustaches, beards, changed hairstyles and glasses. 

Now, some FRT providers say they can even use a source image of a young child and match it against the same person as an adult.

Let’s start with the source (original) image. This needs to be of a certain quality — think about your passport photo and how precise that needs to be — no smiling, a white background, certain light, and certain framing of the head and shoulders. This is all about a consistent and clear base image and sets an “image mapping” benchmark for products like the Australian Department of Immigration’s “Smart Gate”, through to Apple’s “Face ID” to unlock your iPhone.

Modern image mapping technology converts the face data into a unique data sequence using the biometrics of the face. It’s that data sequency which is then stored (not the image). So, when a person walks into a camera frame, the processing computer instantly converts the many faces walking past to data sequences. Those data sequences are quickly searched for against the database of record, and when a match is made, a notification is made.

This is all happening in milliseconds. 

The reality of using FRT gets even more complex when you think about the technical deployment, where the data processing and storage takes place, plus an ever-changing landscape of legislation and permissions (consent) underpinned by the Federal Privacy Act.

Throughout the recent NSW election campaign, FRT was touted as a major advancement in the way problem gamblers, tax evaders and wanted criminals would be detected in venues. There’s a lot in that, as it involves multiple databases across state and federal jurisdictions.

Importantly, the recently-released Club Gaming Code of Practice allows for the use of FRT or digital sign-in, or both. This gives choice to clubs based on size and cost. 

The recent rebuild of the MVSE system will allow this database to be one of the first connections to a club FRT or digital sign-in system, importantly reinforcing responsible gambling as one of the highest priorities for the club industry. 

Alongside this will come the right checks and balances to ensure tech providers, and clubs themselves, protect one of the most sensitive forms of personal data — that of our excluded gamblers.