Facial Recognition (FR) software is a powerful tool when deployed on gaming floors. It can drastically improve a casino’s ability to identify banned and self-excluded individuals, and it can even enhance customer service. Security and surveillance teams that deploy FR are more efficient and can do their jobs more effectively.
But there are still a lot of misconceptions about facial recognition’s efficacy, implementation and capabilities among consumers and new adopters. Omnigo Software is one of the few companies with over two decades of experience implementing FR technology in the gaming and entertainment space, and we’ve heard it all.
Here, we seek to debunk the seven most common myths about FR:
Facial Recognition reduces privacy
In an increasingly digital world, many opponents of FR argue that the technology infringes on people’s privacy. To wit, in 2019, San Francisco became the first city in the U.S. to ban the use of FR by local government agencies. This is a legitimate concern within the context of most public spaces. However, this argument misses the mark when it comes to environments such as casinos and professional gaming facilities.
Everyone who walks into a casino or gaming resort is recorded and surveilled at all times. Patrons expect as much and accept it as part of the security measures required to keep them safe as guests of the property. That’s not to say everyone is being tracked and monitored. Traditional, manual surveillance depends on human judgement and is therefor prone to bias. FR software is incapable of prejudicial judgement and only alerts security staff when the algorithm picks up on individuals who are banned from the property or have placed themselves on self-exclusion lists. FR software actually protects the privacy of casino patrons because it alleviates an opportunity for operator bias. By identifying bad actors who already exist in the database and disregarding everyone else, FR technology essentially acts as a safeguard around privacy.
All cameras can be used with Facial Recognition
While most high-resolution video cameras will work with FR technology, there is a huge benefit in investing in the right cameras. We recommend cameras that are dedicated to face recognition. But it’s not just the quality of the camera that’s important. Camera placement and lighting are integral to the proper functioning of the FR system that’s installed. A facial recognition vendor should work with customers to ensure that cameras are placed in optimal target zones to capture casino patrons from the best angle and that lighting is adequate.
Facial Recognition enhances poor image quality
FR technology has come a long way, but there are still limitations. FR systems can’t just clean up any image to make an identification. For example, if both eyes aren’t visible in a file photo, if the subject is photographed from a side angle, if there are large shadows on the face or if the facial image doesn’t meet the minimum pixel requirements, then the chance of finding a match through FR is lower.
Facial Recognition doesn’t work
Facial recognition is the seemingly simple task of identifying people in photos or video by their face. Humans do it naturally, but it has been a challenge for computers to identify faces for decades. Early iterations of FR used manually generated algorithms that were imperfect and didn’t perform well in complex surveillance environments. However, in 2012, the technology went through a major shift with the successes of deep learning in image recognition. Deep learning is based on artificial neural networks that emulate parts of the human brain. With deep learning models, FR software not only went from achieving human-level abilities, but to exceeding human-level performance within a few years of its inception.
Facial Recognition costs too much work to maintain
Installing a FR system does require some up-front work to ensure cameras are properly placed and that lighting is good, but FR systems significantly reduce the amount of work for security teams in the long run. A good facial recognition system should integrate with your existing database of banned or VIP patrons. This way you should have no additional work after installing the facial recognition solution outside of responding to the face rec alerts.
People are faster than computers at recognizing faces
Studies show that humans can recognize approx. 5,000 facial images and they can only remember a small percentage of names associated with those faces. Casinos entertain millions of patrons per year and have very large lists with people of interest. Today’s FR technology algorithms can easily recognize millions of faces in seconds. Facial recognition that is correctly installed can be used to reduce a large database lookup down to a simple alert showing a possible match of a person entering the facility. In this way, FR is a powerful tool that allows security personnel to focus on appropriate action. It will always be the human operator that makes the final determination whether the alert should be acted on.
People are not comfortable with Facial Recognition
Facial recognition technology is becoming more integrated into our daily lives. It unlocks our phones, tags our posts on social media and validates our identity at ATMs, to name a few of its ubiquitous applications. Like any technology, the more people adopt it, the more culturally accepted it becomes. In time, it will be as second-nature to us as smart phones.
Concerns of individual privacy should never be ignored or downplayed. However, when it comes to current day criticisms about facial recognition technology, some arguments focus on capabilities that do not exist and ignore real-world limitations. That’s not too say that even unrealistic concern and criticism should be discounted. As facial recognition continues to evolve and its applications expand beyond commercial use, the myths and realities of the technology will mature as well. Educating ourselves and our customers not only helps to alleviate unwarranted fears, it helps keep everyone safer.