Built on computer programs that analyze images of human faces to identify them, facial recognition technology has reached incredible maturity in a relatively short period of time. Law enforcement agencies use it to keep society safer, retailers to reduce the incidence of theft, airports to improve the comfort and security of travelers, and cell phone companies to provide customers with new layers of security biometrics.
Facial recognition technology has improved rapidly in recent years due to advances in Artificial Intelligence (AI) and Machine Learning (ML). Thanks to sufficient, high-quality training data and processing power, computers can now detect, classify and recognize faces with high accuracy and speed. Computerized identification of the human face has paved the way for the increasing application of this technology in a wide range of areas.
Following the manifold increase in crimes and terrorist activities in recent decades, the need has been felt for a sophisticated security system for identifying individuals, in which face-based applications such as facial detection and facial recognition can play a role. crucial role.
It is pertinent here to explore how this transformative technology has evolved over the years:
Evolution of facial recognition technique
Early adopters of facial recognition techniques include banks, event managers, forensic investigators, military professionals and law enforcement agencies. Here are some significant milestones in the progress of facial recognition technology over the past few decades:
The origins of this technology date back to the 1960s, when Woodrow Wilson Bledsoe created the first manual measurements of various facial features, including the eyes, nose, hairline and mouth, using electromagnetic pulses, which were entered into a database. Subsequently, upon receiving a new photograph, the system was able to retrieve from the database the image that most resembled that individual.
In the 1970s, researchers Goldstein, Harmon and Lesk established 21 facial measurement points, including lip thickness and hair color, to automatically identify faces.
In the 1980s, Kirby and Sirovich used about 100 facial measurement points using linear algebra, paving the way for a low-dimensional representation of facial images.
In the 1990s, Turk and Pentland invented the first automatic face detection from images. Then the Defense Advanced Research Projects Agency (DARPA) and the National Institute of Standards and Technology (NIST) launched the Facial Recognition Technology (FERET) program, which involved creating a database of facial images . This database was updated in 2003 to include high-resolution, 24-bit color versions of images.
In the 2010s, Facebook created automatic identification of image identities of people whose faces may appear in photos that FB users update daily. Since then, a million photos are uploaded and tagged to Facebook using facial recognition every day.
In 2011, Panama Airport installed a first facial recognition surveillance system to reduce drug smuggling and organized crime. The system resulted in the apprehension of several Interpol suspects.
Significantly, in 2011, the body of al-Qaeda chief Osama bin Laden was identified after he was killed in a US strike using facial recognition. A visual identification was made; Several of his photo comparisons and other facial recognition were used, and a full biometric analysis of facial and body features was done to identify bin Laden.
In 2017, iPhone X became the world's best-selling phone with facial recognition access control used for device security.
Japan currently plans to use facial recognition to verify the identity of athletes at the 2020 Olympic Games, which will be held in Tokyo between July and August.
Steps involved in facial recognition
In the human-computer vision analogy, a camera resembles the human eye and the computer resembles the human brain. The camera captures images of the world and further processing and interpretation is carried out on a computer. Different techniques can be used to process and analyze captured images.
Human can easily identify where one object ends and another begins in a single image. The edges of objects, differences in colors and textures are used for identification on computers, which is called segmentation.
The computer can see any object, perform its identification and also image processing in the same way as human vision, and the appropriate output is also generated. The observed image is interpreted and appropriate analyzes and actions are carried out.
The track of a specific object is found from multiple frames of the image sequence. The speed and distance of the specific object can be calculated as well as future path prediction is performed.
The different points of correspondence between the images, the camera positions and orientation, the reconstruction of the three-dimensional shapes of the scene can be found from several images of a stationary scene.
Broadly speaking, the following steps are involved in facial recognition technology:
Face detection: A computer-aided algorithm pre-processes images captured by the camera . This leads to a clean, easily recognizable image. Subsequently, the localization is performed and the features of the face are extracted.
Facial recognition: The detected facial image is processed and compared with the database of known people; the extracted features are compared with the stored features to find out the probability of a match. They are organized into groups or classes according to similar characteristics to recognize an individual.
Automatic verification of an individual's identity is possible by matching an unknown person's face to the face image in identification records. The identity of an unknown person from the group of people is accomplished by coding it and comparing it with the database that contains coded images of known individuals.
Uses and applications of facial recognition technique
Facial recognition has found its importance in various fields such as security validation and human-computer interaction. As a result, it has become a hot spot for research in the areas of pattern recognition. Researchers are seeking the successful implementation of a functional facial recognition system in the field of computer vision. The exhaustive potential of this technology can be foreseen in areas such as border control and the replacement of key lock mechanisms.
Similar to the Internet, GPS, and many other technologies that have become an integral part of various products today, the roots of facial recognition are firmly planted in the defense and law enforcement sectors, where it is useful in verifying and identifying images.
Law enforcement agencies are widely using facial recognition technology. Cameras at the surveillance site capture images of all objects in real time, continuously and, most importantly, without anyone noticing.
Facial recognition surveillance identifies all individuals as they move through their daily lives. While the ability to recognize individuals in real time has become a reality, in Western countries like the United States, most people in a facial recognition database of documented persons of interest are included due to a history of prior offenses. For example, when retailers want to stop people trying to steal from their stores, they rely on photographs of those individuals uploaded to a private facial recognition database. Since shoplifters are mostly compulsive offenders, an alert can direct store security to watch for these individuals when they re-enter stores. This leads to a reduced incidence of robberies and a much lower chance of violence.
London's Metropolitan Police are using facial recognition technology to find wanted criminals and missing people. The technology was deployed to “specific locations,” each with a “personalized watch list” of wanted people, typically violent criminals.
India is moving towards increasing the use of emerging technologies in law enforcement as police forces across the country plan to replace manual processes with technology-based solutions. Engineering institutes are helping police leverage AI, social media analytics and image processing to identify criminals, manage traffic and prevent terrorist activities.
Police forces in India are beginning to adopt technologies and collaborate with digital leaders to combat crime. They implemented the system of e-filing first information reports (FIRs) to reduce human involvement by submitting them through apps, websites or even an Internet of Things (IoT) device. The comprehensive strategy adopted by the police includes identifying and capturing digital and non-digital evidence using AI and ML, facial recognition and virtual crime prediction. The initiatives led to the construction of a criminal database using AI-based human face detection (ABHED). As officers identify suspicious characters, they must take out their smartphones, click on an image and access the database to determine if he/she is a criminal.
The technology may not be 100% reliable, but it makes the police officer's job easier. It sowed the seeds of a high-tech force that has access to photos, criminal activity and physical details of thousands of criminals at its fingertips. Previously, this data was compiled separately across multiple state districts and different physical records were maintained. All of this made crime prevention an onerous task because the police did not have quick access to it.
Recently, an AI-based prison monitoring system was installed in 70 prisons in Uttar Pradesh. The AI algorithm analyzes hundreds of cameras installed inside prisons to detect violent acts, prison breaches and unauthorized access in real time to timely alert authorities.
There is also little need now to set up police barricades and peer into cars to arrest offenders. Tamil Nadu built an online vehicle base by installing an AI-based system called Tollscope, which was linked to 30 toll plazas. If a vehicle used in a crime crosses any of the squares, the authorities are immediately alerted. Tamil Nadu has also tied up with Telangana police in AI-based automated systems to impose fines on two-wheeler drivers without helmets. The National Crime Records Bureau has begun rolling out an automated facial recognition system across the country to identify criminals, missing people and unclaimed bodies in morgues.
Challenges in applying facial recognition
In the recent past, the field of biometrics has gained most attention due to its reliability for recognition and easy compatibility with available technology. Due to gaps in other identification systems, extensive research is underway in the field of biometrics. Researchers are working towards developing a more user-friendly system to meet the requirements of security systems, which requires more accurate results to protect privacy and assets.
Due to the significant increase in terrorist activities, the importance of a more sophisticated security system for identifying individuals has increased rapidly. Researchers have proposed many approaches to perform human face recognition from images and videos. Following are some of the challenges associated with facial recognition system:
Physical changes: aging; change in facial expression; personal appearance (facial hair, makeup, hairstyle, glasses, disguise).
Geometric changes in the acquisition: Change in scale, rotation of the face in the plane (facing the camera), and location create geometric changes in an acquisition. Rotation, in depth such as obliquely facing camera or profile presentation and unavailability of the full front face, also creates geometric changes in an acquisition.
Image changes: camera variations, lighting variation, channel characteristics (especially in transmitted or compressed images).
Reliability and cost-effectiveness are the main challenging factors for current facial recognition systems. Many researchers target different aspects, such as algorithms that deal with certain problems in facial detection and facial recognition itself. Natural scenes from the real environment may include several factors, such as background noise, variation in lighting conditions, and variation in pose, which may not be present in pre-collected images.
The research also shows that the overall result presents a performance problem in facial recognition. Multiple samples are therefore quite necessary for all techniques. These techniques may fail in special situations such as ID card verification and passport verification as there is only one image for one person.
It is difficult to detect the person from the side view or the image taken from some notable angle. The image of the face with some darkness, partially lighter than the rest of the face, blur, shadows or face with glasses is quite difficult in detecting any individual face.
Despite rapid advances, facial recognition is far from perfect and prone to errors. Even if a facial recognition system has a 99.7% accuracy rate, there is always a risk of catching the wrong person due to the 0.3% error rate. Various environmental factors can affect the accuracy of a facial recognition system. Pertinently, San Francisco and Portland have banned the use of facial recognition on surveillance cameras by public agencies.
The technology has proven controversial, in part because of its ability to invade people's privacy, but also because, without different background data, it may work better for some types of people (particularly white citizens) than others.
Concerns Surrounding the Use of Facial Recognition
There is no doubt that law enforcement depends on the collection of information during the course of the investigation. The instinct of any law enforcement agency to gather information is irresistible and, in fact, is part of its training and standard operating procedures. It is expected that most officers, in their pursuit of public safety, will exercise a reasonable measure of restraint regarding information, and particularly the amount of information that could be collected by facial recognition technology.
There is no harm in using facial recognition for mass surveillance in combination with public video cameras. However, it can also be used in a passive way that does not require knowledge, consent or participation from the subject. The most significant danger occurs when this technology is used for general, suspicionless surveillance systems.
Photographs of citizens possessed by state motor vehicle agencies could be easily combined with public surveillance or other cameras in building a comprehensive identification and tracking system. Any “batch” collection of essentially personal and individual information is a cause for concern. The delicate balance between the “reasonable expectation of privacy” and the ethics of “information as the basis of public security” is not easy to maintain. The availability of phenomenal amounts of information in the absence of any clear plan for its use or purpose makes it responsible for abuses that should concern all citizens.
Unfortunately, facial recognition in authoritarian countries is quickly becoming a routine instrument of police and government control. The practice of police scanning the faces of innocent bystanders for criminals goes against a person's expectation of privacy in public spaces. It's not unlikely that the technology could spread through the network of cameras that eventually cover the streets.
In short
In fact, since the invention of facial recognition in the 1960s, no other technology has sparked more fascination and fear of abuse than the technique of facial recognition. Human rights activists fear that, as a very intrusive surveillance technique, it could offer new opportunities to undermine democracy under the guise of defense.
A balance needs to be struck between our reasonable expectation of privacy and the stated need to collect information that the Constitution provides. The new and ever-growing wave of surveillance technology must not fail the litmus test of personal privacy.
Regarding effectiveness, there is a need to develop an efficient real-time facial recognition system that can work on a system that has a training database of a single image for each individual. Most of the previous research is based on multiple images. Furthermore, most of the work of researchers is based on certain preconditions, which lead to many limitations. As a result of these limitations and performance issues, the practical implementation of facial recognition in a real environment is quite challenging.
These questions lead us to the need to develop more efficient and real-time facial recognition applications and a heterogeneous data matching system with human intervention. Recently, a new approach to hybrid facial recognition technique has been introduced. This facial recognition system is developed with Principle Component Analysis (PCA) and Support Vector Machine (SVM). It is verified against the Oracle Research Laboratory (ORL) database as well as the real environment. With this method based on a single image, facial recognition performance in terms of accuracy is improved. Significantly, Open Source Computer Vision, called OpenCV, has emerged as a library of programming functions. The entire library is cross-platform that offers real-time image processing.
