Biometrics enables the identification and authentication of a person via recognizable and verifiable data in the form of unique and specific biological characteristics. Biometrics can be classified as morphological or biological in type. Morphological identifiers chiefly consist of fingerprints, hand or finger shape, palm vein pattern, the eye (iris and retina), and face shape. For biological analyses, DNA, blood, saliva, or urine may be used by medical and forensics teams.
This automatic user recognition technology uses human physiological and behavioral characteristics with properties like universality, distinctiveness, permanence, and acceptability. There are various methods of biometric authentication and identification, including palm vein recognition, facial recognition, as well as iris and DNA scanning. Retinal scanning, voice dynamics, and handwriting recognition technologies are also in development.
Biometric authentication compares data for a person's characteristics to their biometric template to determine resemblance. The reference model is first recorded and stored. Subsequently, the stored data is compared to the person's biometric data to process authentication.
Biometric identification entails determining the identity of a person using biometric information, such as a photo of their face, a record of their voice, or an image of their fingerprint. First, the biometric data has to be captured, and then compared with the biometric data of other persons in a database.
Palm vein authentication is classified as a method of physiological biometric authentication. It utilizes a vascular image of a person’s palm to record a unique pattern. Palm vein patterns are significantly more difficult for unauthorized users to copy than other biometric features, as they are located inside the body. Since palm veins are mostly invisible to the human eye, they have to be captured under near-infrared (NIR) light.
Palm vein patterns are usually captured under NIR illumination using the reflection method, which entails the emission of NIR light from a person's palm and the capture of reflected light. Because veins absorb much more near-infrared illumination than they emit, the places on the palm where there are veins appear darkened and contrast with the bright areas where they are absent. Thus, this reflection method enables a contactless type of pattern-capturing and user identification can be undertaken.
Palm vein recognition emerged in 1991 and gained popularity as an identification method due to its high security, liveness-detection, user acceptability, and convenience. The palm vein is difficult to imitate because vein patterns are not duplicated in other people, even in monozygotic twins. The vascular patterns on the left and right hands of a person are also different. Palm vein recognition verifies liveness in the presented biometric sample, as the vascular image disappears in the absence of blood flow. Moreover, palm vein recognition systems have high user acceptability, likely in part because they are non-intrusive.
Since these systems acquire data without direct contact with the sensor that extracts the vein pattern image, no contamination occurs between the sensor's surface and the subject's hand. External conditions of the hand being scanned, such as greasiness, the presence of dirt, wear and tear, or dryness and wetness do not affect the scanning process and have no impact on the extracted image.
Palm vein imaging is carried out under the medical spectral window (700∼900 nm), at the near-infrared absorption rate of haemoglobin (including oxygen-containing haemoglobin and deoxidizing haemoglobin) in the palm vein vessels, which is stronger than the near-infrared absorption rate of the surrounding tissues. As a result, when the NIR light is absorbed by the blood vessels, the vascular areas form a shadow. Palm skin consists of three layers: the outermost epidermis, dermis, and innermost hypodermis (subcutaneous), as portrayed on the diagram to the left.
Palm vein images can be extracted using two distinct methods: reflection and transmission. The reflection method lights the target part from the front, whereas the transmission method lights the target part from the back. In the reflection method, the illumination component and the capturing component can be integrated, as they are on the same side. This is different than the transmission method, in which the illumination and capturing parts must face each other. The transmission method needs exceptionally potent light to penetrate the palm skin, meaning that the light source required for this palm vein extraction method is too costly for many applications.
According to Mordor Intelligence, a market research firm, the palm vein biometrics market was valued at $0.58 billion (USD) in 2020. The market research firm projected it to reach $1.86 billion by 2026, following growth at a CAGR of over 22.3% during the forecast period (2021-2026). Another market research firm, Markets and Markets, estimated the size of the global palm vein scanner market to have been $416 million in 2020 and projected it to reach $1,150 million by 2025, at a CAGR of 22.6%.
Markets and Markets cited several factors contributing to the growth, including advantages of palm vein scanners over other biometric technologies, increasing necessity to secure sensitive data for organizations, and growing adoption of biometric identification systems in the BFSI, healthcare, and commercial sectors. The research firm also noted that rising government support for internal biometric techniques in Europe as a result of GDPR compliance efforts as well as an increased number of partnerships and collaborations between vendors in the ecosystem contribute to the growth of the palm vein recognition market.
The Bharat Petroleum Corporation Limited in India is using palm vein recognition technology to keep track of employees and to authorize their access to facilities. Fujitsu announced in 2018 that it will deploy palm vein authentication technology to its employees in Japan. Mordor Intelligence anticipates that palm vein authentication will replace password-based measures for various purposes, such as allowing employees to gain access to buildings and log in to their computers.
The multiplicity of options available for biometric identification enables companies to adopt multifactor authentication and further reduce the probability of fraud. According to the research firm, this and similar trends are expected to make an impact on market growth. For instance, Fujitsu planned to combine palm vein scanning and facial recognition technologies to create a new authentication solution for access control.
Palm vein recognition companies
Some of the advantages of fingerprint scanning include convenience over PIN or password solutions, the secureness of biometrics in comparison to other identity verification methods, and being easier and less expensive to implement than facial recognition. Disadvantages of this method of biometric identification include not being as easy to use as as facial recognition, incapability to take changes in physical characteristics into considerations, and vulnerability to incorrect acceptance and rejections.
Capacitive scanning is the most prevalent type of fingerprint scanning technology. Capacitive scanners use an assemblage of small capacitor circuits that use electrical current to scan and generate an image of the ridges and valleys of a finger. When the biometric data is captured, the processor linked to the scanner analyzes the digital image to identify distinctive and unique characteristics of the fingerprint. There are several advantages of capacitive fingerprint scanners:
- Cost-effective implementation due to economies of scale arising from the availability of outsourced manufacturers
- The increasingly lower cost of implementing the technology behind capacitive scanners makes it a viable option in budget-oriented as well as high-end devices
- A significant benefit of capacitive scanners over optical scanners is that they do not depend on the pattern of light and dark; these scanners instead utilize the physical nuances of fingerprints, which makes them more secure
- The main component of a capacitive scanner is more compact than the component of an optical scanner because it is based chiefly on a semiconductor
- They are faster and more efficient than optical and ultrasonic scanners, making them the fastest fingerprint scanning technology
- Capacitive scanners can be installed on physical buttons or solid surfaces. Other input gestures, as well as swiping and scrolling, can also be integrated along with the fingerprint-reading function
A disadvantage of capacitive fingerprint readers is their unsuitability to be used as in-display fingerprint scanners because the technology is incompatible with the capacitive touch input technology used in display technologies such as IPS, LCD, or OLED. Manufacturers of mid-level to high-end devices are moving away from capacitive fingerprint scanning because of a design trend favoring more screen space and bezel-less screen edges. Its common implementations include placing them alongside a physical button or the accessible surfaces of devices.
Optical fingerprint sensors (also known as optical scanners) are frequently used in implementations of in-display fingerprint scanning. The technology entails the capture of a two-dimensional optical image of a fingerprint and analysis of the particular pattern of its ridges and valleys. A light sensor system called charge-coupled device (CCD) is an integral part of optical scanning technology. CCDs are also used in camcorders and digital cameras and operate by utilizing LEDs to illuminate the surface of a finger and enable the sensor to capture a detailed image of it. The advantages of optical scanners include the following:
- They can be integrated with a capacitive display screen, therefore enabling in-display fingerprint scanning and allowing manufacturers to create devices with more screen space and minimal bezels
- The implementation of optical fingerprint scanning technology is not as expensive as implementing ultrasonic fingerprint scanners, enabling manufacturers to use the former in mid-level devices
- The entire system is considerably small; the total area of an optical module can measure less than one millimeter
There are some limitations of optical scanners:
- Most CCD components are unable to always distinguish between a picture of a finger and the finger itself
- Optical scanners capture a two-dimensional image, meaning that prosthetics and low-quality images can be employed to bypass the protection
- In general, optical fingerprint scanning technology is not as secure and reliable as the technologies used in capacitive scanners and ultrasonic scanners
- Optical scanners are slower than capacitive scanners, and newer ultrasonic scanners are significantly faster
Ultrasonic scanners are the most novel of the three fingerprint scanning technology types and use an ultrasonic transmitter and receiver to obtain fingerprint data. These scanners generate and transmit an ultrasonic pulse against the finger. The particular ridges, valleys, pores, and other characteristics of a finger cause part of the pulse to be absorbed while the rest is reflected back to the sensor, which maps out a three-dimensional image of the fingerprint. The advantages of ultrasonic scanners include the following:
- The ability to capture a high-fidelity three-dimensional image of the fingerprint
- They are more secure than capacitive scanners and more reliable than optical scanners
- An ultrasonic scanning module can be used as an in-display fingerprint scanning solution, therefore offering compatibility with designs with a lot of screen space and bezel-less edges
- Capability to operate under non-optimal conditions, such as if the users have dirty or wet hands. Unlike capacitive scanners, they do not depend on electrical charges
There are limitations of ultrasonic scanners:
- The underlying operational mechanism makes them slower than capacitive scanners.
- In-display fingerprint scanners may be incompatible with certain screen protectors because the ultrasonic pulse cannot penetrate through thick surfaces.
Fingerprint authentication companies
Facial recognition is a method of identifying or confirming a person's identity by scanning the characteristics of their face. Facial recognition systems can be used to identify people in photos, videos, or in real time. This technology is mostly used by security and law enforcement agencies, but is increasingly adopted in other areas, such as to easily and securely log in users to devices. In such use cases, facial recognition does not require a large database of photos to determine the identity of a person, instead identifying and recognizing a single individual as the sole owner of the device and limiting access to others.
Aside from being used to unlock users' phones and computers, facial recognition can match the faces of people present in areas viewed by specialized cameras with the images of individuals on a watch list. Watch lists can contain pictures of anyone (also people who are not suspected of having committed a crime) and the images can be sourced from anywhere, for example social media. While facial technology systems can differ in type, generally they work in the following way:
- Face detection: The camera first detects and locates the image of a face, either in isolation or among other people, showing the person facing the lens or in profile.
- Face analysis: Subsequently, an image of the face is captured and analyzed. The majority of facial recognition technology relies on 2D rather than 3D images, as it it easier to match a 2D image with public or private database photos. Specialized software then reads the face's geometry, noting such factors as the distance between eyes, the depth of eye sockets, forehead to chin distance, the shape of cheekbones, and the outline of lips, ears, and chin. The objective of this is to identify unique facial characteristics that can serve as the key to distinguishing the face.
- Conversion of the image to data: The face capture process transforms the analog information taken from a face into a set of digital information (data) that corresponds to its features. Through this process, the data harvested from the face is essentially converted into a mathematical formula, and the resultant numerical code is called a faceprint. Just as thumbprints are unique and specific to a person, every individual has a distinctive faceprint that can be utilized biometrically.
- Finding a match: As the final step, the faceprint is compared with other recorded faces in a database. For instance, the FBI has access to up to 650 million photos, collected from various state databases around the US. Facebook and other social media platforms where users upload images of themselves can also be used as databases. In the case of Facebook, any photo tagged with a person’s name is integrated into the platform's database, which can be referenced by facial recognition systems. If a faceprint successfully matches an image in a facial recognition database, then a determination is made.
Of all the biometric measurements, facial recognition is considered the most natural. Intuitively, this makes sense, since we typically recognize ourselves and others by looking at faces, rather than thumbprints and irises. According to estimates, more than half of the world's population is in some way affected by or interacts with facial recognition technology on a regular basis.
Facial recognition companies
Iris scanning biometrics measure the unique patterns in the colored part of an eye (iris) to verify and authenticate a person's identity. Biometric iris recognition is contactless, fast, accurate, and can operate at long distances, with some solutions requiring only a glance from a user. Unlike software-based biometric solutions such as face and voice recognition, iris-based identification requires specialized hardware, making it the less common choice in consumer-oriented markets. However, owing to innovations in biometrics that have made the technology more accessible in terms of cost and installation, iris biometrics has grown in popularity across the vertical markets and also in the consumer electronics sphere.
Iris recognition systems create images of the eye using infrared light. Because melanin in the iris is transparent upon exposure to infrared illumination, iris detail can be recorded irrespective of eye color. Recognition algorithms locate the boundaries of the iris and then process the portion of the image containing it to generate a specific and focused representation of the iris' pattern. This representation provides a very high level of differentiation between individuals in a population.
Presentation attacks necessitate capturing an image of an individual’s iris with enough detail to enable the creation of an appropriate artifact. To maximize the effectiveness of a presentation attack, the task of capturing an iris image can be carried out with an iris recognition camera and the cooperation of a subject, although successful attacks have been undertaken with a general purpose camera. Early iris recognition systems were vulnerable to simple presentation attacks that used photographs of eyes, but more sophisticated features in certain modern iris systems featuring liveness detection make this significantly more difficult.
Iris recognition is considered to have a very low false match rate and is often the chosen biometric identification method in cases where population sizes are large. Locating and isolating the iris from larger images of the eye is a challenging task that can result in false non-match errors if executed unsuccessfully. Patterned contact lenses also lead to false non-matches, restricting the usefulness of iris recognition on populations using them. There are two use cases for iris recognition; close-up and at a distance. For close-up iris recognition, hardware to illuminate the eye region and capture the iris image is relatively straightforward to implement, while the recognition of the iris image at a distance is harder. Where imaging distances range from one to five meters, iris recognition systems can be implemented using readily available, commercial-grade equipment.
Iris recognition is often used as a physical access control solution in high-throughput environments that benefit from speed and accuracy. It is also commonly utilized in border control, as a method of identifying travelers as they pass through borders by land, sea, and air. Iris scanners have also been integrated in consumer mobile devices, a development that has led to the expectation of the rise of iris recognition in FinTech, IoT, and other vertical markets among market forecasters. Samsung’s Galaxy Note7 and Microsoft's Lumia 950 and Lumia 950 XL Windows phones all feature iris scanners, as well as multiple editions of the Fujitsu Arrows devices marketed by NTT Docomo, a mobile carrier specializing in the development and implementation of smartphone iris biometrics.
Iris recognition companies
Whereas biometric authentication and identification methods such as fingerprint, face recognition, and iris scanning are based on the measurement of similarity of external feature points that are in constant evolution, the scanning of deoxyribonucleic acid (DNA) is intrinsically digital as it remains unaltered throughout an individual's life and even after their death. While DNA's polymorphic information content (PIC) is potentially useful as a biometric identifier, offering advantages such as accuracy, strictness, discriminatory power (and ease of expanding this power), and the ability to use one unified analysis platform all around the world, it is not yet widely used in this way.
DNA can be obtained from various biological sources, such as body fluid, nails, hair, and used razors. For biometric applications, a buccal swab is the simplest, most convenient, and least intrusive sample collection method. Buccal cell collection entails rubbing a small piece of filter paper or a cotton swab against the inside of a person's cheek in order to collect shed epithelial cells. Subsequently, the swab is air dried or pressed against a treated collection card to transfer the epithelial cells for storage purposes.
The most prominent drawback is that DNA analysis is time-consuming compared to other authentication methods: it takes at least four hours to obtain short tandem repeat (STR) identification data by common methods used in forensic science. PCR amplification and electrophoresis take up the most time of the DNA analysis process, and it is impossible to condense the duration of these steps with existing technologies. Single-nucleotide polymorphism (SNP) analysis is faster, offering the possibility of analyzing up to 96 SNPs in a 30-minute time frame, and an SNP-based system could be used in passport identification or in very large-scale trade transactions; however, its utility is limited.
The polymorphic target region in DNA used to create the DNA ID does not correspond to a person’s physical characteristics or disease factors, since the STRs and the SNP samples are picked from the extragenic regions. Nevertheless, because the DNA ID system involves managing sensitive information that can identify a person, the process should be carried out under rigorous supervision to protect the privacy of the affected individuals. Once the DNA ID is generated, the one-way encryption makes the recovery of any of the original DNA information impossible, meaning that particularly the raw materials, such as the buccal swab, should be strictly controlled to prevent spoofing.
A large amount of capital is needed for DNA analysis to buy and maintain equipment and for commercial kits. In addition, a laboratory and a group of molecular biology specialists are necessary. Although these high costs may be an obstacle to venture capital investment, it is expected that with the increasing popularity of DNA techniques, the unit costs of the apparatus and reagents will be lowered.