To date, the number of smartphone (smartphone) users worldwide has exceeded 30 billion and is projected to grow by hundreds of millions over the next few years.The countries with the highest number of smartphone users are China, India, and the United States, and each country has a user mark of well over 1.2 million.Although the population ratio is different, you can see how many people are using it compared to Japan's population of XNUMX million.
In this way, smartphones that are closely related to daily life are being pursued for usability, and their functions and performance are further improving.In addition to the technological innovation of the equipment itself, with the advent of 5G (evolution of communication technology) that enables high speed, large capacity, low latency, and multiple simultaneous connections, there is a possibility that services related to our lives will further evolve.
This time, we will pick up biometric authentication (biometrics authentication) and "face authentication", which have high usability and are adopted in smartphones in the connection between "people's lives" and "information".
table of contents
What is biometric authentication?
There are three main types of technology for identifying a person: "knowledge authentication," "property authentication," and "biometrics."Among them, "biometric authentication" is a technology that authenticates an individual based on physical and behavioral characteristics that are characteristic of a person but never have the same characteristics as other people.There are various types such as fingerprint, face (whole), iris (eye), ear shape (auricle), voice print, vein, palm print, and behavioral biometrics.
Among them, "face recognition" has a wide range of uses, and is reflected in various potential applications such as infrastructure, security for providing each service, and embedded applications.Of course, there are privacy issues and security concerns, but it has the potential to be integrated in any place where smart devices and the latest cameras are installed.
"Face recognition" is the future development of the information society (digital society), how to quickly and accurately recognize a specific individual and respond to the content and nature of procedures and services that diversify and expand with the development of Society 5.0. It may be said that it is one key in.
About the ideal personal authentication platform with an eye on Society5.0 (Ministry of Internal Affairs and Communications)
https://www.soumu.go.jp/main_content/000560721.pdf
Face recognition (What is Facial Recognition)?
Facial recognition system is a branch of computer vision.It consists of "face detection" that finds information about a person's "face" from digital images such as surveillance cameras, and "face recognition" that reads desired information from face data.
Face recognition by AI software is the mainstream.Various information obtained from the image is converted into an array of points by a biometric algorithm and then replaced with numbers.Based on that number, it is subdivided into, for example, light and dark → people → eyes, nose, mouth, etc.A large amount of data stored in the computer in advance is patterned by machine learning, and the solution of the condition obtained by collating with the data is returned.By setting a model case (requirement) in advance, it can be diverted to various services.
It is said that it is difficult to distinguish under some conditions including twins, and the accuracy of face recognition is not yet perfect.However, it is expected that the identification rate will improve further in the future as AI machine learning, which has the identification ability equivalent to that of humans, advances.
Survey by the National Institute of Standards and Technology (NIST) (FRVT, face recognition vendor test)
The system developer (vendor) voluntarily submits a face recognition algorithm to NIST for testing, which is useful for the Face Recognition Vendor Test (FRVT). On November 2020, 11, we released the face recognition accuracy by face mask.It can be seen that the accuracy of face recognition of each vendor is improving year by year.
FRVT Face Mask Effects
https://pages.nist.gov/frvt/html/frvt_facemask.html
What kind of purpose is it used for?
Face recognition is said to be somewhat superior in terms of security compared to the risk of forgetting, losing, or stealing passwords, IDs, credit cards, etc., because the characteristics of the "user" are subject to discrimination.In addition, it is considered to be an advantage that it is "non-contact biometric authentication" that does not burden the user mentally and can be easily introduced into consumer devices.As a familiar example, it is used for unlocking and payment on mobile devices such as smartphones, signing in to mobile apps, entry / exit management of companies and hotels, and events such as concerts.
Soft infrastructure related
In October 2020, Singapore was the first country to introduce facial recognition as an ID card.By integrating "face recognition" into the authentication system of the login ID "SingPass (Singapore version My Number)" required for electronic transactions with government agencies, it is possible to perform personal identification such as final tax returns with "face". became.Currently, limited services such as pension (CPF), public housing (HDB), NS (military service) are used in each app using biometrics or passcodes.
Security related
Law enforcement agencies around the world also use biometric software to scan faces in CCTV footage to identify people.At the airport, we are expanding the operation of "face recognition gates" and using them to confirm the identities of travelers.
Incorporating such facial recognition technology into surveillance has become commonplace in the last decade, with 10 countries permitting its use for surveillance purposes.By incorporating a computer program that analyzes and identifies images of human faces, it is possible to find not only criminals but also lost children.
Current status of monitoring by face recognition (reference)
The Surfshark website publishes data on the status of facial recognition technologies in 194 countries.Get a better understanding of the current state of face recognition monitoring around the world.
Mapped: The State of Facial Recognition Around The World
https://surfshark.com/facial-recognition-map
Coronavirus infection countermeasures
Face recognition does more than just identify a particular individual.Due to the recent pandemic of new coronavirus infection, wearing a mask has become "new normal", but face recognition technology has been incorporated to identify "wearing a mask".
It is used for measures such as checking the wearing of the mask via the camera installed at the entrance of the store and calling out if the wearing is not grasped.Not limited to wearing masks, there is a possibility that specific conditions may be set, so expectations are high for responding to diversifying services.
Other services
As for detailed needs, there is an increasing demand for personal authentication in the new normal "mask-wearing" state, as opposed to "persons who are not wearing masks" as a countermeasure against coronavirus infection.There are also services that incorporate face recognition technology into health care, from entry / exit management to physical condition management.
In addition, the data for face recognition collected by passing in front of the camera also contains information that can grasp consumer behavior, and demand is expected to grow in the retail marketing industry.It may be wonderful to connect people, things, and things if we can quickly obtain consumer information and provide services that meet the needs of users.
Face recognition issue "spoofing"
There is a cyber attack called "spoofing". In "face recognition", it is said that it occurs at the time of registration or authentication, and it is considered that countermeasures are necessary.When considering the introduction of a face recognition system, it is a good idea to check what kind of measures are necessary.
Types of spoofing
2D paper photos and digital images, high resolution video, paper masks with cutouts in eyes and mouth, special masks like those used in Hollywood, wax figures, realistic dolls, video projections, sleep with eyes closed Users, look-alikes, Doppelgenger, etc.
In addition to the above, it is also important to take measures against spoofing attacks (Presentation Attacks) against face recognition methods (including biometric authentication devices), and research to automatically detect these is also becoming active.
At FaceTec, a 3D face recognition provider,Protect your personal information with standardized biometrics presentation attack detection (PAD) tests (non-sharing, non-phishing, virtually spoofing prevention, etc.).The white paper explains the detection and importance of standardized spoofing attacks in biometric systems, as well as the implications of being certified as an anti-spoofing standard for facial recognition technology.
Standardized Testing for Biometrics (reference)
https://www.zoomlogin.com/FaceTec_PAD_Testing_Whitepaper.pdf
Cameras and software for face recognition
Face Recognition Solution Development Kit 3DiVi Face Recognition SDK
API and software development kit for face recognition solutions using 3DiVi's fast and highly accurate algorithms.Process videos and photos in real time to detect, analyze and search human faces.
Main features
-Detect, analyze and search human faces
-Real-time video or photo processing
-On-premises, cloud, edge, or embedded processes
The algorithm developed by 3DiVi has been evaluated as the highest ranked algorithm in the world in the face recognition technology benchmark test conducted by NIST, and the recognition performance when wearing a mask is also very high.Please check the video for details.
Intel RealSense ID Camera
Intel RealSense ID Is a camera specializing in "face recognition" released by Intel in January 2021. The RealSense ID series combines a depth sensor and a special neural network to provide the security, privacy, reliability required for face recognition, and excellent authentication speed in a more compact housing than a credit card. ..
Secure and accurate facial recognition is useful for a variety of applications for everyone, anywhere (eg smart locks, point-of-sale kiosks, ATMs, vending machines, access control, etc.).
* This solution supports children to tall adults (55 cm from the camera installation point and 120 cm to 190 cm in height).In addition, it is possible to authenticate from darkness to strong sunlight and can be used indoors and outdoors.
* Intel RealSense ID Solution F450, F455 Datasheet
https://dev.intelrealsense.com/docs/intel-realsense-id-f450-module-and-f455-peripheral
Face recognition cameraIntel RealSense ID CameraCan be used as soon as it arrivesF455 Peripherals (Kit with camera body, cable, tripod, etc.) and for embeddedF450 ModuleThere are two lineups.We are currently pre-ordering and the manufacturer is planning to ship in March.We will handle it, so please feel free to contact us. ■
* Added on March 2021, 3:
RealSense ID F455 Camera Peripheral is in stock now.In stock and immediate delivery.
Vzense DCAM face recognition ToF depth camera
Vzense Technology's ToF (Time-of-Flight) depth camera has excellent "speed" by reducing the workload of the computer, "safety" that is optimal for protecting PII (personal identification information), and excellent. It has features such as "accuracy" due to distance accuracy.It is useful for face recognition and behavior analysis (number counting system).
Example of application (face recognition) using Vzense
– Face ID authentication
Get 3D face data with ToF depth camera for accurate and efficient live identification
– Smart home
Accurate face recognition and lifestyle patterns using 3D information can increase the level of security protection for door locks
– Access control gate
Incorporating 3D facial recognition technology enables highly safe and stable access gate machines in critical locations and intelligent control with high efficiency without human intervention.
– Security monitoring
By monitoring the movement trajectory of people throughout the day and grasping the dynamic information in real time, security is strengthened more efficiently and quickly.
LIPSedge ToF depth camera
It is a ToF depth camera designed for various indoor / outdoor applications such as AR / VR, games, machine vision, medical detection, face recognition, indoor XNUMXD measurement, and mounting on drones and automobiles. It is also compatible with multiple platforms such as Windows, Linux and Android, and can be applied to industrial and enterprise applications.
And with Intel's RealSense technology and the OpenVINO toolkit, like the LIPSFace AC1, which has a versatile flip design concept for seamless conversion of desktops, wall mounts, stands, and turnstiles in one unibody, it takes about 770 seconds. There are also models that achieve ultra-fast, non-contact authentication.
■Intel RealSense Camera D400 Series
Features such as face recognition are realized by combining the RealSense SDK 400 for Intel's RealSense D2.0 series with other software platforms such as OpenCV and OpenVINO. SDK 2.0 is constantly being updated, so "middleware" solutions may emerge in the future.
Introduction to the Intel Distribution of OpenVINO toolkit with the Intel RealSense Viewer
https://www.intelrealsense.com/realsense-openvino/
LIPS Face ™ AI Facial Recognition (face recognition system) provides SDK (SW3) for 500D single-face recognition and SDK (SW2) for 200D multi-face recognition. You can select according to.
Summary
Industries that require biometrics range from security, medicine, banking and public sector services, marketing research, and personal identification."Face recognition" was taken up this time, but the attention of "biometric authentication technology" is very high, and even in the future and trends of biometric authentication technology in 2021 (announced on November 2020, 11), the daily life of people all over the world It has been announced that some are becoming more rapid.Through integration with mobile devices, many of us interact with some form of biometrics every day.It can be said that we are looking forward to seeing what kind of relationship will be created in our lives in the future.
The Future of Biometric Technology and Trends in 2021
https://recfaces.com/articles/biometric-trends
Our company (Tegara Corporation) handles ToF cameras and depth cameras from one.
We also accept consultations on mass production.
Even if you do not know the distribution channel in Japan, we will investigate and guide you whether it can be handled.
Please feel free to contact us if you have any requests such as "I want this software / hardware sold by this manufacturer!".
