Yearly Recap: Biometrics and Digital Identity 2023

January 3rd 2024

An overview of the pivotal developments in the biometrics & digital identity realm throughout the year 2023.

article image

The year 2023 has been marked by significant developments in the biometrics and digital identity landscape, showcasing ongoing technological advancements, widespread adoption across various sectors, and the emergence of regulatory measures to tackle new challenges.

Major tech giants such as Apple, Google, and Microsoft have continued their support for passkeys on their platforms and applications. Notably, platforms like X, TikTok, LinkedIn, and GitHub initiated their passkey rollouts during the year. WhatsApp also introduced a new feature enabling the locking of private chats using biometric authentication.

In the regulatory realm, there has been a notable focus on setting standards for AI and the metaverse. The European Union’s EUBI Wallet underwent pilot programs, and the global implementation of decentralized identities gained momentum.

Intel introduced its FakeCatcher technology, contributing to the ongoing efforts to enhance security in digital identities. ID R&D made strides by integrating voice recognition capabilities into ChatGPT, adding a layer of biometric security.

Breakthroughs in biometric technology extended beyond traditional methods, with notable progress in DNA capture, brain biometrics, and biometric identification within virtual reality environments. Additionally, the promising accuracy of heartbeat-based biometric verification emerged as a noteworthy development in the field.

Mainly, 2023 has proven to be a dynamic year for biometrics and digital identity, marked by technological innovations, regulatory advancements, and the expansion of biometric applications into new and exciting domains.

Without further ado, let’s dive into some key highlights of the year:

January — February

Research:

  • Researchers develop fingerprint biometrics and liveness for under QLED smartphone displays: A group of South Korean researchers has created a sophisticated biometric recognition and liveness detection system using a multi-factor approach and sensors integrated beneath quantum dot light-emitting diode (QLED) displays. Their research paper details the integration of fingerprint biometrics with the analysis of skin temperature variations.
  • Haptic authentication for blind, low-vision users tested by US, and Canadian academics: Researchers from the University of Waterloo and the Rochester Institute of Technology have experimented with an inventive authentication method for blind and low-vision (BLV) users. Named OneButtonPIN, this approach utilizes haptic vibrations to enhance the user-friendliness and security of authentication for individuals with BLV.
  • The uniqueness of fingerprints from birth explained in academic study: A group of researchers has presented conclusive evidence in the journal Cell, establishing that fingerprints exhibit unique biometric characteristics from birth. The published paper explores the formation of fingerprint ridges, revealing them as epithelial structures guided by a spatial pattern regulated by a Turing reaction-diffusion system. The development process parallels that of hair follicles before shaping into distinctive biometric features. According to the researchers, prenatal molecular and cellular mechanisms govern the formation of these structures.
  • A study reveals motion data can identify people in VR: A study suggests that the distinct movements of an individual’s head and hands within a virtual reality (VR) application can serve as a unique identifier. The research, based on data collected from tens of thousands of users engaged in the VR rhythm game Beat Saber, reveals that users can be uniquely identified across multiple sessions by analyzing their head and hand motions relative to virtual objects. The study, which included data from individuals using the VR application developed by Beat Games (acquired by Meta in 2019), highlights that biomechanics can act as a distinctive identifier in VR. With just 5 minutes of data per person, the study achieved a unique identification accuracy of 94.33 percent from 100 seconds of motion and 73.20 percent from only 10 seconds of movement, showcasing the potential of biomechanics as a comparable identifier to more commonly used biometric modalities like facial and fingerprint recognition.
  • Clinical research shows infant fingerprint biometrics nearing real-world effectiveness: The confirmation of newborns’ identities in healthcare settings using biometrics appears to be attainable with current technology, as suggested by a paper published by the U.S. National Library of Medicine.
  • AI put in charge of creating fairness training of AI: Researchers claim to have devised a method to establish data “pipelines” for furnishing training models with high-quality, artificially generated faces. However, their findings indicate that, in this specific scenario, artificial intelligence does not outperform debiasing models when it comes to training AI.

 

Industry news and developments:

 

March — May

Research:

  • Breakthrough made in DNA capture: Researchers have showcased the capability to capture human genetic data, including medical and ancestry information, from minuscule fragments of DNA found in the environment using environmental DNA (eDNA) techniques. This development has prompted concerns related to privacy, legality, and ethics, as the captured data could be potentially used for efficient surveillance of individuals with specific ancestral backgrounds, medical conditions, or disabilities. The emergence of eDNA technology underscores the importance of implementing comprehensive genetic privacy regulations to address the ethical and privacy implications associated with such advancements.
  • Stanford study showed AI benchmarks aging poorly, and need work: An AI research team led by Stanford University has determined that algorithms are surpassing humans in certain benchmark tests. However, the study also reveals that AI benchmarks, as a whole, are aging and becoming less effective. The team suggests the need for new and redesigned benchmarks. This insight is particularly relevant to biometric systems as businesses, governments, and universities expand the use and capabilities of algorithms, including those for facial recognition.
  • Another idea for biometric authentication — and two threats: A U.S. researcher, Chen Wang from Louisiana State University, claims to have developed an algorithm that utilizes hand and behavioral biometrics for identity authentication. This innovation is designed to address the growing vulnerabilities observed in some existing biometric security measures. Wang aims to enhance device security without relying on traditional biometric scanning methods, which are becoming increasingly susceptible to attacks. According to LSU, Wang’s approach involves using a person’s hand as a sounding board. In this method, a service seeking to authenticate an individual on a phone would transmit largely inaudible ultrasound signals in narrow frequencies between 17 and 22 kHz, causing the phone to vibrate in the person’s hand.
  • Biometric anti-spoofing handbook updated with liveness competitions, legislative impact:The Handbook of Biometric Anti-Spoofing has been updated with the publication of its third edition by Springer. This edition aims to offer comprehensive and authoritative guidance on presentation attack detection (PAD). The updated handbook provides broader coverage of PAD methods across various biometric modalities, such as face, fingerprint, iris, voice, vein, and signature recognition. Additionally, the third edition includes information on major PAD competitions, valuable databases for researchers, and an analysis of the impact of recent legislation on biometric anti-spoofing.

 

Industry news and developments:

 

June — August

Research:

  • Researchers defeated voice biometric security by targeting common liveness approaches: Researchers from the University of Waterloo, a pair of Canadian computer scientists, have identified shortcomings in voice presentation attack detection, making it surprisingly easy to bypass certain biometric defenses. In their experiments to circumvent voice authentication, the researchers reported a success rate of up to 99 percent in only six attempts against the least effective biometric security system they tested. These findings have been published in the IEEE Computer Society’s digital library.
  • AI vs. AI: MIT researchers combat image manipulation: A team from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) has designed a new tool to jam AI image generators, using invisible “perturbations” at the pixel level of an image. A release describes how the PhotoGuard technique uses a combination of offensive and defensive tactics to block AI tools such as DALL-E or Midjourney from manipulating photos to create deepfakes and other compromised images. In the encoding tactic, perturbations are small alterations to the latent representation of an image that an AI engine “sees” in mathematical code. By making changes to the code, PhotoGuard “immunizes” the image by making it incomprehensible to AI, which can then only perceive it as a random entity. The resulting output will be unrealistic and recognizably altered — faces on a grey field, for instance, or unblended into a blurred background. On a defensive level, PhotoGuard creates perturbations in the original input image that are checked against during the inference process, which causes the AI to confuse the two images. This more complex biometric “diffusion attack” uses significantly more memory than encoding.
  • Coughers identified in algorithm study: Researchers from the University of Washington and Google have developed code that is reportedly adept at identifying individuals within a group based on their coughing patterns. This technology could be applied in diagnosing and treating conditions like cystic fibrosis, tuberculosis, and asthma. Additionally, it might prove effective in identifying populations affected by future forms of respiratory illnesses such as COVID-19. According to the researchers, their model correctly classifies coughing sounds 82.15 percent of the time. In a departure from traditional methods, the researchers used multi-task learning, with the second task being biometric speaker verification among four subjects from a dataset of natural, real-world coughing instances.
  • Researchers revealed PVC pipe spoof threat to voice biometrics: Digital security engineers at the University of Wisconsin-Madison have discovered a weakness in automatic speaker identification systems that can be exploited using PVC pipes readily available at most hardware stores.
  • AI-powered ‘DyLiN’ System Can Turn 2D Images Into Dynamic 3D Models: Carnegie Mellon University (CMU) and Fujitsu researchers have pioneered a method of converting a 2D image into a three-dimensional one. This is a challenging task because it requires understanding how to handle deformations and changes that occur when going from 2D to 3D. The researchers’ Dynamic Light Field Network (DyLiN) method uses artificial intelligence techniques to tackle this difficulty. It learns a deformation field, which is a way to describe how the shape of an object changes, from the input images to a canonical representation of the object. This canonical representation helps to handle any abrupt changes or discontinuities in the object’s shape.

 

Industry news and developments:

 

September — December

Research:

  • Researchers developed artificial ‘fingerprints’ that can be reset if breached: Researchers have developed an artificial micro-fingerprint generator that utilizes cholesteric liquid crystals (CLCs) to create cryptographic credentials resembling fingerprints. These credentials can be reset in case of data compromise. Cholesteric liquid crystals can self-assemble into intricate patterns with unique visual properties, forming physically unclonable functions (PUFs). These PUFs can function as cryptographic keys or artificial fingerprints that are challenging to replicate. The generator, when subjected to a high-frequency electric field, reorients microdroplets of a chiral liquid crystal containing a photoluminescent dye, creating a distinctive pattern resembling a fingerprint. Besides resets, this technology could enhance anti-counterfeiting features like holograms and RFID tags for increased security.
  • Experiments show promise for real-time fingerprint liveness detection: Researchers from China, Portugal, and Canada have developed lightweight software for real-time fingerprint biometric liveness detection. The method they propose requires shorter training and fewer parameters compared to previous approaches. The researchers utilize a broad learning system (BLS) for fingerprint liveness detection, a novel approach that doesn’t require GPU training. BLS is designed to enhance the performance and applicability of detection algorithms on mobile devices. The proposed method includes a three-step process, and authentication can be carried out following positive liveness detection.
  • Facial recognition can get better at spotting people from a distance: A study published in Nature addresses the challenge of optimizing surveillance cameras for facial recognition at a distance. Many surveillance cameras with facial recognition capabilities are limited in their range. The study proposes a facial recognition method based on deep learning using a dataset that includes information such as the distance of individuals from the camera, the focal length of image sensors, and the size of the target face in pixels. The extended dataset was created using the Georgia Tech Face and Quality Dataset for Distance Faces. Testing the method on various surveillance camera image sensors revealed that several sensors achieved an average accuracy above 99 percent in the recognition process. The researchers suggest that this approach could be crucial for security applications in smart cities.
  • NYU research combined subject and age sets to improve aged-face imaging: Researchers at New York University’s Tandon Engineering School have developed a latent diffusion model that can age images of people’s faces while preserving genuine facial identifiers. This technology could have various applications, including identifying missing persons who have aged over the years, realistically augmenting actors’ faces in movies, and potentially spoofing remote facial age estimation systems. The latent diffusion model is a sophisticated method that allows for the transformation of facial images with age progression while retaining important facial features.
  • Assessing skin hue to skin tone makes measuring biometric bias easier: Researchers from Sony and Tokyo University have developed a new method for measuring apparent skin color in computer vision, aiming to address concerns about biases in facial recognition algorithms. In their pre-print paper on Cornell University’s arXiv server, the team introduces a “simple, yet effective, first step towards a multidimensional skin color score.” They have incorporated a skin hue axis, ranging from red to yellow, in addition to the traditional skin tone, creating a multidimensional color scale. This x-y chart combines both tone and hue to assess fairness in algorithms, providing a more nuanced and comprehensive evaluation of apparent skin color.
  • MSU researchers developed whole-body biometric recognition system with $12M grant: Researchers at Michigan State University have developed technology for long-range biometric identification, supported by a $12 million four-year federal grant from the Intelligence Advanced Research Projects Activity (IARPA) under its Biometric Recognition and Identification at Altitude and Range (BRIAR) program. The goal of the program is to create end-to-end software systems capable of detecting individuals and extracting “biometric signatures” from the entire body, including gait and body shape, in addition to facial features for multimodal biometric matching. Drones are utilized in this system to recognize individuals at a distance, taking advantage of increased altitude and range for imaging. However, the vantage point often captures the tops of people’s heads, limiting the reliance on facial recognition alone.
  • Study claims success, inspires app using face biometrics to match dating pairs: A study published in the journal IEEE Access suggests that facial recognition software can “reliably predict” a person’s personality, achieving an accuracy rate of at least 70%. Conducted by five Chinese researchers, the study used an algorithm to classify static images of individuals based on the big five personality traits: neuroticism, extroversion, openness, agreeableness, and conscientiousness. The accuracy of the algorithm was compared against subjects’ self-reported personality traits. The research was based on a dataset of 13,347 pairs of face images and personality traits, with a deep neural network used to test the leftover biometric samples against self-reported traits.
  • Brain biometrics helped identify sports concussions: Novel brain biometrics could help inform whether an athlete is ready to return to play following a concussion, according to new research from the University of South Australia. Conducted in partnership with the University of California San Francisco (UCFC), researchers found that changes in micromovements of the brain — termed ‘headpulses’ — could detect the lasting impacts of a concussion. Using a custom-designed headset to evaluate headpulse biometrics among 101 amateur male and female Australian Rules Football players in South Australia, researchers identified brain abnormalities in 81% of players inflicted by concussion, signalling sustained injury beyond expected recovery times.
  • Researchers proposed a method for hiding faces while allowing biometric searches: A research paper presented a proposed universal face obfuscation method for a family of commercial off-the-shelf (COTS) facial recognition algorithms. It argues that evolving privacy laws, such as GDPR, necessitate obfuscation in facial recognition systems, and the proposed method aims to obscure faces from human eyes while still allowing recognition by biometric algorithms.
  • Heartbeat accuracy for biometric verification showed promise and limitations: A study explored a novel biometric verification method called CompaRR, which captures beat-to-beat information embedded in intervals between heartbeats. While many existing biometric verification methods involve stationary physiological signals like facial authentication and fingerprint scanning, non-stationary signals like heartbeats are considered more difficult to fake. Traditional ECG (or “EKG”) recordings have demonstrated the ability to verify identity through the heart’s electrical activity, but these methods are typically complex and require medical professionals to setup. CompaRR aims to use a more cost-effective and user-friendly approach by measuring beat-to-beat time intervals, making it suitable for applications with wristbands and video cameras.
  • Tiny radar biometric sensor could identify people by their heartbeat: A seemingly impossibly small radar sensor capable of detecting biometrics including heartbeats has been created by researchers at the University of California, Davis. The sensor is described as being the size of a sesame seed and funded by the Foundation for Food & Agriculture Research. While that seems random, the battery-powered sensor, which was built in the school’s Electrical and Computer Engineering Department, is so sensitive, that it can hear a small leaf thinning as it runs out of water.

 

Industry news and developments:

  • X launches account verification based on government ID: X, formerly Twitter, has launched government ID-based account verification for paid users to prevent impersonation and give them benefits such as “prioritized support.” The social network has partnered with Israel-based Au10tix for identity verification solutions. The pop-up for ID verification indicates that the Au10tix could store this data for up to 30 days.
  • ID R&D brings voice recognition to ChatGPT: ID R&D has been showcasing a new user authentication solution specifically designed for ChatGPT at an industry conference this week. IDVoice for ChatGPT leverages the company’s voice recognition technology to verify an end user during a speech-enabled ChatGPT session, in which a user can vocally interact with OpenAI’s famous chatbot.
  • Meta suggests an AI computer vision fairness standard and opens a model to all: Meta made a couple of significant computer vision announcements last week. It introduced a proposed fairness benchmark, and it made the vision model open source. In both cases, the parent of Facebook wants to insinuate itself deeper into the fabric of AI development. Meta has proposed FACET as the standard for image classification and semantic segmentation “at an unprecedented scale.” How much, if at all, Facebook benefits from this is an open question. The company famously swore off facial recognition for the social media service.
  • Patently Apple reported Apple has filed a new patent application for a new Face ID system designed for smart glasses and, potentially, the Vision Pro. This innovation would allow Face ID to synchronize with the cameras on and in the head-mounted display (HMD) to unlock companion devices such as Apple Watches, iPhones, iPads and Mac desktop computers.
  • Amazon announced passkey rollout for biometric passwordless login: Following a quiet rollout, Amazon has announced that it supports passkeys on browsers and the iOS Amazon Shopping App.
  • Amazon’s One palm biometrics readers for businesses got a lukewarm introduction: Amazon’s cloud computing subsidiary AWS (Amazon Web Services) has lifted the lid on a new palm-scanning identity service that allows companies to authenticate people when entering physical premises.
  • BIO-key connected with AWS sales reps: BIO-key has joined the Amazon Web Services Independent Software Vendor Accelerate Program, which is meant to connect vendors offering software the runs on AWS with the AWS Sales organization, including its globally dispersed field sellers.
  • Mastercard joined Mercedes to put finger-sensor purchases within cars: Global finance company Mastercard says its mobile pay systems have been integrated with Mercedes-Benz vehicles in a service called Mercedes pay+. In Germany, 3,600 gas stations also are being integrated.
  • Google’s Pixel smartphone, the Pixel 8, has been unveiled, featuring “Class 3” facial recognition technology: That’s a considerable improvement over the Pixel 7’s “Class 1” technology, which was too weak to be used to authorize app logins or payments — applications that the Pixel 8 will allow.
  • Microsoft introduced a new authentication recommendation engine for Entra: Microsoft was not entirely happy with how companies are using its identity and access management product line for enterprise clients Microsoft Entra. The tech giant has come up with a new solution. Customers of Microsoft Entra will be automatically enrolled into Microsoft Entra Conditional Access so-called “intelligent policy engine.” The engine gives recommendations on security settings. The idea is to help customers figure out how to have more granular control over authentication and access.
  • Fime Asia expanded Visa accreditation for contactless payment terminal testing: Fime Asia in Taiwan can now validate Level 2 compliance for the contactless payment terminals of vendors in Visa’s network after the French company received accreditation from the payment cards giant. For Level 2 transactions, businesses collect and process additional data for each transaction from the buyer. Elements such as the customer code, tax amount and tax identification are collected and processed along with Level 1 data such as credit card number and expiration, billing address, and zip code.
  • NFL’s Tennessee Titans and Verizon to use face biometrics for fan access control: The National Football League’s Tennessee Titans and Verizon have announced they are partnering to verify guest identities using facial authentication for secure access through the 5G Edge Accelerated Access opt-in system along with dozens of 5G Ultra Wideband cell sites at the Nissan Stadium this season.
  • Idemia asserted top results for biometric performance across modalities in NIST tests: The results of testing by the U.S. National Institute of Standards and Technology show top-ranking accuracy for Idemia algorithms across iris, fingerprint, and face biometric modalities, the company says in an announcement. The update to NIST’s IREX 10 benchmark showed the company’s iris biometrics algorithms deliver the most accurate match results for single-eye comparisons. The same algorithm, submitted in June, sits third on the leaderboard for two-eye accuracy.
  • Aware added biometric capabilities, ease of use, and development upgrades to the platform: Aware has enhanced its face biometrics platform with a range of new functionalities and launched a new developer hub to ease the adoption of biometrics by businesses.
  • VirtualSignature-ID selected FaceTec: UK-based and government-certified digital signature provider VirtualSignature-ID has selected FaceTec to build selfie biometrics and 3D liveness detection into its platform to help small and medium-sized enterprises adopt digital identity verification.
  • Civic introduced Proof of Personhood with FaceTec biometrics and liveness: On-chain digital ID service provider Civic announced Proof of Personhood for decentralized apps (or “dApps”) building on Solana to verify users’ identities, screen out AI, and protect their services from abuse.

 

 

The year 2023 witnessed a dynamic landscape of innovation and adoption in biometrics and digital identity, with ongoing efforts to address security challenges, enhance user experiences, and navigate the evolving regulatory environment. The expansion of passkey authentication, global digital identity initiatives, and advancements in AI and biometric technologies marked key trends throughout the year.

Read ‘Biometrics biweekly’ on the Paradigm Platform.

Subscribe to Paradigm!

Medium, Twitter, Telegram, Telegram Chat, LinkedIn, and Reddit.

 

Main sources

Research articles

Biometric Update

Science Daily

Identity Week

Find Biometrics

Tags

biometrics
tech