Facial Recognition Technology (FRT) is seeing exponential growth. This recent market report reveals that the global facial recognition market reached US$4.29 billion in 2020, and is expected to reach US$13.87 billion by 2028 with a CAGR of 15.7%. FRT market revenue growth is primarily driven by an increased need for physical security, the rise in deployment of biometric devices, and the expansion of 5G technology.
This article highlights some key considerations for choosing a facial recognition software solution, and why FaceMe, CyberLink’s facial recognition technology, is the best fit for IoT/AIoT hardware integration. We will also demonstrate its use in various scenarios such as security, surveillance, and fintech.
Edge computing has become a core technology thanks to its role in IoT devices and networks. The hardware and chipsets on which FRT engines run have dramatically improved in power, speed, cost, and size. In addition, software has become more accurate, eliminating issues of bias, gaining powerful new features, and increasing in flexibility. Considering all these elements, it’s no surprise that some of the world’s largest brands have adopted facial recognition software.
Here are the critical considerations in choosing the best facial recognition software:
CyberLink’s FaceMe AI Facial Recognition is a cross-platform SDK that provides a 99.73% accuracy rate. It is ranked top-tier in the NIST (National Institute of Science and Technology) FRVT (Facial Recognition Vendor Test). FaceMe provides major features such as face detection, face compare, face search, and face match. It is optimized for multiple operating systems and mainstream IoT/AIoT hardware, chipsets, and SoCs.
For more information, please visit the FaceMe website or refer to our guide on the 7 Success Factors for Choosing the Best Facial Recognition Solution.
Founded 25 years ago, CyberLink is uniquely equipped to pioneer facial recognition technology. A longtime leader in creative editing software, CyberLink has spent decades developing advanced digital media editing features, including recognizing faces in photos with a high degree of accuracy. With over 200 patents and a vast team of specialized engineers, this expertise and precise technology became the foundation for CyberLink’s facial recognition solution, FaceMe.
CyberLink strives to be a leader in the ethical use of facial recognition. Developing, distributing, and innovating its technology in accordance with Global Human Rights Principles, the company’s priority is to respect all internationally recognized human rights, and to contribute to positive human rights impacts wherever possible. For these reasons, FaceMe is not sold or used in any block-listed countries or for any use cases which may infringe upon human rights.
FaceMe is constantly updating to integrate new, cutting-edge features and industry advancements. Maintaining close partnerships with top chipset makers, Original Equipment Manufacturers (OEMs), and Original Design Manufacturers (ODMs), CyberLink continuously updates FaceMe to make the most of the latest hardware, optimizing speed, accuracy, and energy use.
FaceMe is designed to maximize recognition accuracy without sacrificing flexibility. Other facial verification solutions often lack the ability to operate in multiple form factors, on any operating system, and with a variety of chipsets and hardware.
The National Institute of Standards & Technology (NIST) is the preeminent authority on the accuracy of facial recognition systems. Their tests measure the accuracy of solutions, weighing characteristics differently depending on use case.
FaceMe reached up to 99.73% accuracy in a 1:N VISA BORDER test in Investigation Mode at 1E-6 accuracy (meaning the False Match Rate, or FMR, is 1 in 1,000,000), making it appropriate for use cases like border checks and financial security identification. In the 1:1 WILD test, ideal for security and surveillance use cases, FaceMe scored a strong 97.00% at IE-5 accuracy (FMR: 1 in 100,000). Both tests ranked FaceMe in the top 10 of all vendors.
The facial recognition engine from FaceMe includes a comprehensive set of precise features making it suitable for a wide range of use cases, from smart retail and surveillance, to access control, to financial security (electronic Know Your Customer), smart city, and more.
Face detection is the first step in facial recognition. FaceMe is capable of concurrently detecting multiple faces from different angles and counting the total number of faces. Upon receiving images, FaceMe will begin a face search, which is the process of searching the frame for a facial template. Simultaneously, it will recognize and count the number of live individuals in the frame.
Simply put, 1:1 face matching and comparing analyzes the similarity of two faces to verify if they are the same person or not, i.e., a match. 1:1 face match is frequently used, such as to unlock your phone with Face ID, or to verify that you have access to financial services. All data is encrypted using HTTPS and TSL. Most importantly, FaceMe does not save the actual face image.
The 1:N face search means verifying a person's identity by comparing their face against a certain number (“N”) of faces in the database. The most common use of face search is an access control system to verify that a person in front of the camera works at the company – if so, their face was pre-enrolled in the company’s facial database and will be recognized.
The facial attribute feature pulls valuable data on the characteristics of the faces analyzed. Predictive details like gender, age, facial expression (happy, sad, angry, surprised, etc.), and head orientation can provide crucial analysis and insights to improve customer experience and planning.
FaceMe’s anti-spoofing feature supports both 2D and 3D depth cameras. It ensures that the image being processed is that of a live person and not a 3D printed mask, video, photo, or other spoofing technique. The anti-spoofing technology from FaceMe greatly reduces the risk of identity theft, of particular importance in the fintech industry.
The International Conference on Computer Vision (ICCV) conducted its Face Anti-Spoofing Challenge at ICCV 2021. It ranked FaceMe in the top three algorithms globally and in first place when excluding vendors from China and Russia.
Face mask recognition features have become a must-have in facial recognition systems today. Thanks to earlier work in medical use cases where face masks are standard, FaceMe has long had a specialized facial recognition feature adapted to perform with high accuracy on masked individuals. FaceMe can now provide mask detection, checking that individuals are wearing masks and wearing them correctly, covering their nose and mouth. It simultaneously executes highly accurate facial recognition on masked individuals, maintaining up to 98% accuracy in these settings.
All these tasks can even be completed in crowded settings as the algorithm can instantly scan all faces in a frame, making it ideal for busy public spaces such as airports, shopping malls, and concert venues.
FaceMe’s facial recognition features include face detection (recognizing and counting live individuals in frame), template extraction (extracting data about a face in the frame), face search (searching the frame for a known face template), and face matching (comparing all faces in a video to an existing template to find a match). To learn more about FaceMe, you can visit the FaceMe website.
FaceMe outperforms competitors in flexibility and is designed to be easily integrated into form factors as small as a door lock, unique as a vehicle, or common as a typical security workstation. Offered as both a software development kit (SDK) and plug-and-play package like FaceMe Security, it can be quickly installed on various operating systems and optimized for mainstream CPUs, GPUs, VPUs, and SoCs.
Factors such as power, accuracy, speed, cost, and scalability can indicate the best infrastructure for each use case. FaceMe can run on various configurations, from servers and workstations to IoT/AIoT devices and mobile phones. Find out the most optimal hardware for installing FaceMe.
Additionally, FaceMe supports a wide range of operating systems – Windows, Linux, Android, iOS, and JetPack – assuring developers that they can easily integrate facial recognition technology on their hardware platforms, even if equipped with different OSes.
CyberLink leverages partnerships with the most reputable names in chipsets, hardware, software, and distribution to ensure FaceMe remains the most optimized solution available on the market.
Chipset partners like Intel, NVIDIA, Qualcomm, Mediatek, and NXP work with CyberLink to ensure every version of FaceMe is optimized to run their newest, strongest, and most widely available technology. Leading names in hardware like Advantech, Supermicro, VIVOTEK, BKAV, and ELO design the world’s best hardware and collaborate with CyberLink to determine the most efficient and powerful setups for running FaceMe.
Software partners like Milestone and NX Witness ensure FaceMe is compatible with, and seamlessly integrated into, existing security and video management software.
ACEBiotek needed a facial recognition solution to help with employee clock-in and health management at its factory of over 20,000 employees and contractors. Previously, the company used access control cards but was looking to upgrade to protect against card swapping issues and create a better health management system in light of the pandemic.
FaceMe was integrated directly into the company’s Wallie Screen Access Control and Health Screening System. Now, when workers arrive at the employee entrance, they are automatically identified, clocked in, and checked for mask compliance and temperature, all in less than one second. This new process ensures the facility is secure and COVID-safe and eliminates the bottlenecks and long queues that plagued the previous process.
A large tech company incorporated FaceMe to provide access control and security on its campuses.
FaceMe performs employee authentication to ensure no unauthorized individuals gain access to the campus, simultaneously minimizing internal security risks by ensuring employees only gain access to areas of campus relevant to their roles and clearance. The company can access records of employee movement when needed, and employees never have to worry about a lost or stolen ID card.
A large tech company integrated FaceMe into its smart city offerings for public safety and services.
FaceMe’s robust AI security monitoring and access control tools met many city needs such as video surveillance, facility access control, and incident management. The benefits of FaceMe extended to the city’s entire population, giving residents the option to enroll in facial recognition powered services, including accessing their public services accounts, contactless admittance to public transit, and skipping lines when entering city facilities like museums and parks.
Japanese digital key platform Bitkey provides smart locks, smart access, and security solutions for homes and buildings. Bitkey needed a reliable and highly flexible facial recognition solution to integrate into its smart door locks.
FaceMe’s ability to be integrated into a form factor as small as an individual door lock made it the best option for Bitkey. Today, Bitkey’s smart locks powered by FaceMe provide residents with the highest level of identity verification for accessing their homes, all in a swift, contactless experience.
FaceMe is available as an SDK or as several plug-and-play packages, including FaceMe Security, featuring health and VMR add-on solutions, and FaceMe eKYC & Fintech.
The FaceMe SDK is a cross-platform facial recognition engine ideal for custom scenarios and use cases where you want complete control to mold the algorithm to your unique software infrastructure. It is optimized to run on most hardware configurations, from high-end workstations to low-power chipsets typically used in IoT devices, and compatible with Windows, Linux, Android, and iOS systems. To learn more about the technical details, read the FaceMe SDK system requirements.
The FaceMe SDK includes all the advanced face recognition, attribution, image pre-processing, and mask features described in the features section above. These specialized features, such as face search, gender, and age attribution data, anti-spoofing, and mask detection, make the FaceMe SDK ideal for smart security, public safety, smart retail, and smart banking applications. Market-leading accuracy as tested by NIST and ICCV, combined with optimized performance and high flexibility, make the FaceMe SDK the most comprehensive and reliable SDK on the market.
FaceMe Security is a turnkey, value-added facial recognition solution that runs on industrial PCs, workstations, servers, and more. It can integrate into video management systems, IP cameras, door locks, and time and attendance software. It can conduct a wide range of tasks from identity verification to visitor check-in/employee clock-in, mask detection, real-time monitoring, and notifications.
As a plug-and-play software package, FaceMe Security is a perfect option for well-defined use cases and quicker timelines. It best accommodates typical security applications and is highly scalable, meaning it can be used with a single camera or even multi-camera, multi-location deployments. For more information, you can read about the FaceMe Security’s system requirements.
The FaceMe VMR add-on offers live multi-camera monitoring along with video recording and playback, offering competitive value for individual deployments or at scale. Optimized to minimize processing power for each camera feed, it remains flexible enough to integrate into edge, cloud, or hybrid environments. It offers 24/7 real-time monitoring of multiple, simultaneous streams, with either gallery or floor plan views.
FaceMe Security’s Health add-on conducts three crucial checks: 1) mask detection, including whether the mask is being worn properly, covering the nose and mouth; 2) identity verification; and 3) temperature detection, using a compatible thermal camera to flag temperatures above normal. Employers, retailers, schools, and other facilities can utilize the Health add-on to ensure individuals remain healthy and to reduce virus transmission as much as possible.
FaceMe eKYC & Fintech is a turnkey facial recognition software package with unique biometric verification and identity validation features relevant to the Banking, Financial Services, and Insurance (BFSI) industry. Easily deployed at the edge or in the cloud, the technology offers financial institutions the ability to deploy eKYC solutions remotely while protecting against biometric fraud.
Compared to other identity verification methods, such as fingerprinting and iris scanning, facial recognition is the most accurate, efficient, and remotely deployable. Anti-spoofing features protect against fraud techniques such as using a photo or mask to fool the identity verification mechanism.
FaceMe eKYC & Fintech is ideal for applications such as opening a bank account remotely, cardless ATM transactions, mobile banking, and insurance services. As with other FaceMe solutions, FaceMe eKYC & Fintech is a highly flexible software that can run on various operating systems and hardware setups. For more information review the system requirements for FaceMe eKYC & Fintech.
There is no doubt that facial recognition will continue to grow in popularity as the most accurate type of biometric technology, powering the most efficient security processes available today. With more players quickly coming to market, it can be challenging to choose the best facial recognition solution for you.
Certain factors, such as accuracy, are a critical consideration regardless of use case. You must have a reputably tested and highly accurate solution to use facial recognition ethically. Next, consider your unique use case and weigh factors like flexibility, feature sets, cost, and more.
Be sure to do thorough research on the history and background of the company behind the software you choose. How long have they been around? Where does their expertise in facial recognition come from? Do they have ties to countries where data privacy may be a concern?
Facial recognition is powerful technology, so the vendor and software you select must be tested, trusted, and reputable.
People who read this article also checked out:
7 Success Factors for Choosing the Best Facial Recognition Solution for Your Unique Needs. There are many different factors to consider when choosing a facial recognition solution. Reading this article will ensure you can successfully implement your solution.
We're here to help!
Our team of experts will be happy to answer your questions and schedule a demo. Start building your own facial recognition solution today!