Facial recognition helps airports tackle access control head on

The days of physical access control solutions (PACS) simply consisting of key cards and personal identification numbers (PINs) are long gone. Nowadays, to better secure access to restricted areas, incorporating biometrics such as facial recognition into a PACS is allowing airports to not only increase security, but also efficiency.

The overall facial recognition market is estimated to grow from US$3.4 billion in 2016 to $6.8 billion by 2021, at a CAGR of 15.3 percent from 2016 to 2021, according to a report by MarketsandMarkets. The major forces driving the market are the growing surveillance market, increasing government deployment and increasing applications in numerous industry sectors.

The number of attacks on airports worldwide has made increasing security measures a priority for many governments. By deploying facial recognition technology in airport PACS, security operators are better able to secure and restrict access to high-security areas, while also making it easier to identify “unwanted” persons.

Main uses for facial recognition in airport PACS The need to strengthen airport security against unauthorized access is just one reason to deploy facial recognition technology. According to Alexander Khanin, Founder and CEO of Russia-based VisionLabs, the main use cases for face recognition in PACS at airport terminals are as follows:

• Identification and real-time tracking of all passengers and personnel entering airport terminals. This includes terrorist and “wanted” list person identification; blacklists of an airport, airline or any other organizations; and verification of passengers using face and photo ID data;
• Passenger quantity control in airport terminal zones;
• Passenger database clusterization (frequent flyers data analysis service for the airline companies);
• Automated and semi-automated passenger check-in;
• Unauthorized access prevention to the security check zone during flight check-in;
• Unauthorized access prevention to the passport control zone;
• Automated or semi-automated border control based on e-gate with face recognition as one of the identification subsystems;
• Identification and tracking of all passengers and personnel in the airport safety zone;
• Gate access control and automated people counting.

Airports are also using real-time facial recognition technology for surveillance applications. “The technology compares faces seen by the camera to one or multiple image databases and instantly finds known individuals,” said Elke Oberg, Marketing Manager of Cognitec Systems.

“Airports can detect and prevent unwanted behavior in much faster and more efficient ways, as security agents can track individuals in real time, or receive alerts on mobile devices to act within the immediate vicinity of a suspect. Based on the anonymous analysis of faces seen by the camera, security staff can also receive an alarm if too many people gather in a specific area, measure waiting times to direct traffic or detect if a person does not pass through a high-security area within a required time frame,” she added.
 

Facial recognition as part of multifactor authentication

Facial recognition in a PACS works best when it is used in combination with other layers of security. “Airports are using face recognition as one biometric technology for authorized access to high-security areas, usually in combination with another token or biometric,” Oberg said. “Adding a biometric identifier lowers the risk of an unauthorized person gaining access with a stolen token.”

Roger Rodriguez, Director of Business Development at Vigilant Solutions, elaborated on this point, saying, “When a door or gate entry access is monitored and tracked with the use of biometric technology, it offers an even greater level of security by providing any organization the ability to lock and unlock doors with multifactor authentication.”

He pointed out how before the introduction of biometric technologies, many organizations relied on traditional identification cards or employee badges for authorized entry and access. “History has shown these cards present vulnerabilities in security as they can easily be fraudulently replicated, intentionally handed off to others, lost or stolen.

However, when these cards are combined with a biometric-like facial recognition technology, the additional layer of video surveillance access control offers extra protection and eliminates the fraudulent access seen with traditional ID card systems. This added technology makes for a reliable integrated security system,” he said.
 

Machine and deep learning propel facial recognition technology

Machine learning has played a big role in improving the accuracy of many different technologies across various spaces in recent years. When it comes to facial recognition technology, machine learning is playing an equally important role.

“Accuracy and speed of facial recognition algorithms are benefiting from deep learning methods, therefore increasing the reliability and efficiency of the technology in access control solutions,” Oberg said.

Rodriguez pointed to the non-intrusive nature of facial recognition technology as the reason it benefits from machine learning. “What makes facial recognition as a biometric stand apart from other identity-based recognition solutions such as iris scanning or fingerprints is that it is non-intrusive. It requires no physical contact and access control systems can now grant access even from certain distances. ‘Faces’ can be captured without a person’s knowledge, making the machine learning technology more robust,” he said.

Rodriguez explained biometric sciences, such as facial recognition technology, rely on computer systems to validate and authenticate specific measurable physical characteristics unique to an individual. As such, this technology has proven itself to help establish identities and access, making machine learning widely accepted in modern day society. Furthermore, since machine learning has made significant advancements in accuracy and reliability, the demand for PACS has grown in many areas including law enforcement, government and many commercial spaces.

Similarly, Gary James, Head of Sales and Marketing at Aurora, pointed to artificial intelligence, specifically deep learning techniques, as a “game changer” for his company’s facial recognition systems. Deep learning techniques help “by delivering significant improvements in every aspect of the biometric operation,” he said, which has improved the accuracy and customer experience exponentially.
 

Lighting challenges in airports

In terms of technical challenges, James noted how ambient lighting conditions on image capture could negatively impact the results of facial recognition. “This is a particular problem at airports where large ticket halls and other areas often have very large window areas, so sunshine (or the lack of it) means that ordinary video surveillance-type cameras sometimes cannot deliver a good enough image for reliable face recognition,” he explained.

To overcome this Aurora uses specialist IR sensors and illuminators designed specifically for this task. “The result is a system which operates consistently with all faces and in all conditions,” he said.
 

Overcoming privacy concerns

Privacy concerns are a major challenge facing facial recognition technology in PACS. Biometric information is very personal and the idea of one’s face being stored in a database somewhere is disturbing to many. These privacy concerns have been reignited recently with several major airlines around the world deploying new facial recognition procedures for passenger boarding.

Rodriguez acknowledged “the widespread use of machine learning systems like facial recognition have caused many privacy concerns across the United States and other parts of the world.”

Many of these concerns, however, are the result of a lack of education and over-dramatization of the usage of facial recognition in Hollywood. “The reality here is the alarmists are simply misinformed and many of their claims remain untrue. The benefits of biometrics like facial recognition and machine learning offer identity solutions which ensure public safety and even empower businesses to better understand their customers and workplace,” Rodriguez explained.

Companies like Vigilant Solutions are working with lobbyists in the U.S. to educate both sides on best practice for biometric facial recognition technology use. “Vigilant Solutions evangelizes transparency, accountability, a need for systemwide monitoring and metrics, and encourages all our customers to implement policy before facial recognition deployments are made or mandated by the courts,” Rodriguez said.
 

No stopping facial recognition

Despite any concerns over privacy, the benefits of facial recognition technology in airport security cannot be denied, especially in a time where ensuring passenger safety, as well as the safety of all those that are employed at an airport, is of the utmost importance. Taking the necessary steps to ensure such security, though, is unlikely to come without controversy. In airports, we will continue to see more and more deploy facial recognition in PACS, as well as for passenger boarding and customs and boarder control.