As we enter a new decade, facial recognition is a hot topic. Media outlets frequently report on the technology as a crime-solving potential or a privacy concern. While both claims are valid, the real story is harder to uncover.
Many of us use facial recognition every day as a tool to unlock our cellphone or to tag photos. But the truth is that the use of facial recognition goes far beyond social media. From local police to border patrol agents, the technology assists law-enforcement in their duty to protect and serve. Many facial recognition systems require a training process that is critical to the success of this technology.
The adoption of facial recognition (FR) has sparked public debate regarding its current use and the risk of future abuses. Civic groups and politicians question the intrusion of FR into everyday life and have concerns about the lack of accountability. For over ten years, federal, state, and local law enforcement have used facial recognition to combat sophisticated criminal activity. The steady march of technological advancement in FR invites scrutiny and heightens the fear of privacy loss.
Requests to ban facial recognition technology may stem from a misunderstanding of how FR works and conflicting information about the limits of the technology. Facial recognition works by verifying the identity of a person using their face. It captures, analyzes, and compares patterns based on facial details. Then, the technology transforms it into digital data by applying an algorithm and compares the image to those held in a database to generate a faceprint, unique to a single individual. A faceprint allows for greater security and privacy as it mathematically maps an individual’s facial features.
According to the Security Industry Association (SIA), a trade association for global security solution providers, most of the confusion surrounding facial recognition concerns government use, particularly by law enforcement. The use of facial recognition in the U.S. is guided by established policies and procedures, addressing many of the public concerns. SIA “believes the technology should be used in a safe, accurate and effective way, and is working with the U.S. Congress to help set the example on how to ethically and responsibly govern this technology."
Myths around Facial Recognition
Facial recognition is rampant in the United States, with little regard for privacy. In the United States, the use of biometric data is determined by established laws and rules to ensure privacy. Government use of these technologies has been restrained by the U.S. Constitution with specific safeguards to protect free speech and religious choice, citizen’s rights to due process, and unreasonable searches and seizures.
Facial recognition matches are often wrong and implicate innocent people. Despite media reports about the implications of misidentification, facial recognition systems do not decide the identity of a person. Facial recognition is just another tool that law enforcement uses to assist in identifying a person of interest. Today, courts do not consider search results as evidence – the search only offers FR as another investigative tool that may or may not provide value.
How to select a facial recognition software solution partner
Facial recognition is considered the most natural of all biometric measurements. We recognize ourselves by looking at our faces, not by looking at fingerprints or irises. Keep in mind the following questions when choosing a facial recognition partner:
What is the difference between a standalone facial recognition system (FRS) engine vs. a complete system?
There are many standalone facial recognition engines on the market. But their downside is that they require a considerable amount of time and effort to integrate into a production system. The chosen FR system should easily integrate with any existing VMS systems in place. Also, the FR system should be available as a full integration platform with supported video recording, access control, and fire alarm integration capabilities.
What are the lighting requirements for the system to perform at the advertised recognition rates?
The FR system requires a camera to detect a face and capture a snapshot. The snapshot must be high quality so that the system can accurately compare it to the face images already stored in the database. Although many factors determine the quality of the snapshot, lighting is especially important. Most system challenges are due to scenes that are too dark, bright, backlit, or susceptible to changing lig