Facial Recognition

What Is facial recognition system?

Facial recognition system is a technology used to detect the presence of a person by comparing a digital image or video of a person’s face against pre-existing data. Facial recognition and facial comparison can be used to verify a person’s identity by recording and analyzing an image or video of a person’s facial structure and comparing it against a pre-existing image to determine whether there is a match. The pre-existing image can be from a private or public database or the image on a government-issued ID.

Facial recognition systems have been in development for decades, but advances in the last few years have made these solutions commonplace in our daily lives. Facial recognition tools are now available in smartphones and financial institutions are beginning to adopt facial comparison for digital identity verification as part of digital account opening.  

Here we will provide a high-level overview of the technology, its benefits, drawbacks, and where it might be utilized in the future.
 

How facial recognition Works

Different facial recognition tools work in different ways, but in general, they follow a simple four-step process:

Step 1: Face Detection

Whether alone or in a crowd and from either a video or still image, the camera detects a human face.
 

Step 2: Analysis

The technology performs a detailed analysis in real-time on the facial image using machine learning or artificial intelligence by assessing the location of 80 nodal points on the human face. The arrangement of these nodal points is unique to the individual, and the data set represents facial features such as the distance between a person’s eyes, the bridge of their nose, and the curvature of their cheekbones.
 

Step 3: Converting an Image to Data

The unique arrangement of nodal points is then converted into biometric data usable by facial recognition and facial comparison system. The resulting numerical code is called a faceprint.
 

Step 4: Making a Match

Once the faceprint is established, this biometric data can be compared against other sources of data, such as existing faceprints in a public or private database, or the image on a government-issued ID document. If the system discovers a match, it can notify whatever application is making use of the technology.
 

What is facial recognition used for?

Facial recognition technology is proving valuable in a variety of use cases for a variety of organizations. Here are a few of the ways facial recognition is being used for effective, efficient identity verification today:

• Device Security: An individual’s face can be used as an authentication factor to unlock their mobile devices. In this use case, the live image is compared against a pre-existing image on the phone’s private database. Select models of both Apple iPhone and Microsoft Android mobile devices are equipped with this capability.
   
• Anti-Theft Prevention Measures: Law enforcement agencies can use the face recognition systems to identify suspects on known criminal databases after a theft has occurred.   
   
• Purchasing Alcohol: Some bars and liquor stores use facial recognition tech to detect fake IDs and driver’s licenses while facilitating a better customer experience. For example, with facial recognition technology, customers who consent to their data being kept on a private database owned by the store can securely purchase alcohol at a self-checkout counter.
   
• School Security: A facial recognition system can be used in a school to identify known suspicious persons such as expelled students, drug dealers, or other threats. The administration can then contact school security to address the situation.
   
• Airport Security: Airports are already using facial recognition to expedite the security and boarding process. In some countries manual ID document checks can be performed by e-gates which use facial recognition and facial comparison to verify a travelers ID document. Furthermore, government agencies have used the technology to identify individuals who have overstayed their visas or are under a criminal investigation.
   
• Law Enforcement: Using facial recognition technology, law enforcement can significantly increase their efficiency to threats like identity fraud. For example, the New York Department of Motor Vehicles has used a facial recognition program to identify 21,000 potential identity fraud instances since 2010.

 

Benefits & drawbacks involved in the use of face recognition

As is the case with many emerging technologies, facial recognition systems have both advantages and drawbacks to their use:

Benefits of Facial Recognition

• Analyzing Crowds: A law enforcement office can wear a body camera equipped with facial recognition technology to analyze the biometric information of a large crowd of people, compare the information to a facial recognition database of mugshots, and spot potentially dangerous attendees, such as those on a terrorist watch list. This task would be nearly impossible without the assistance of facial recognition technology.
   
• Customer Experience: Facial recognition as an authentication factor provides a superior customer experience to passwords and other biometric factors, such as thumbprint.
   
• Digital Account Opening: With the use of facial recognition technology, financial institutions can make mobile apps more secure and build secure account opening processes that can be completed digitally without ever stepping foot into a bank branch. Watch this brief video below to see how facial recognition fits into the digital account opening process.  

Drawbacks of Facial Recognition

• Privacy concerns: Facial recognition technology has inherent privacy concerns associated with its use. It is fully capable of tracking an individual, and some like the American Civil Liberties Union (ACLU) have expressed concern over potential violations of civil liberties, civil rights, or human rights should the biometric technology be used by police departments, the FBI, Department of Homeland Security, or other law enforcement. For this reason, some cities in the United States, such as Oakland and San Francisco in California and Portland, Oregon, have instituted facial recognition bans for their police departments. In addition, the Commercial Facial Recognition Privacy Act of 2019 was introduced to the United States Congress which seeks to regulate how private companies and other entities can collect, process, and store facial recognition data.
  
  Furthermore, large scale data breaches are not uncommon. Privacy and security advocates are uneasy about the possibility of a data breach leaking this biometric information onto the Dark Web. 
   
• Data Misuse: An extension of the privacy issues are concerns over data misuse. A Pew Research Center study found that respondents were uncomfortable with facial recognition technology being used as surveillance technology to monitor apartment buildings, track attendance of employees, or to analyze how people respond to advertisements. This stems from a fear that private companies and governments will misuse this data.
   
• Identification Errors: Facial recognition systems are not perfect. If there are flaws in the image used to build the faceprint, the system may not be able to match it with another in the facial recognition database. Poor resolution, heavy shadows, the angle of the face can all distort the final faceprint and cause false positives. In addition, a small or incomplete facial recognition database may create similar errors from the other end. Though its accuracy has increased significantly in recent years, there are still challenges facing biometric facial recognition software.
   

The future of facial recognition software

The future of facial recognition is uncertain. There is already much debate and disagreement as to how this technology should be used. At the same time, non-governmental organizations are adopting the technology more and more. It is widespread and becoming more so. The conversation is not over, but it is highly unlikely the door will be completely closed on facial recognition technology. The National Institute of Standards and Technology (NIST) is expanding the conversation as well with new research, including the effects of race, age, and sex on facial recognition systems as well as the effect of face masks related to the coronavirus. The continued study of the technology will lead to a more informed conversation around its potential value and ethical concerns.