iCatch’s SoCs are powering a diversity of camera applications in
iCatch’s Ci1 intelligent ISP Solution Enhances Reliability and Safety of Face Recognition in Access Control Systems
As the rapid advance of security technologies over the past few years, the smart access control devices have migrated from simple pin code verification or card key sensing to biometric authentication. The fingerprint recognition is the mainstream solution for biometric sensing applications nowadays. On the other hand, face recognition caught some new attention recently in certain areas, such as hotel guest room or residential house entrance access, fleet vehicle driver authentication, and payment authentication in a bank or shopping. Access management by face recognition can improve the convenience of usage and increase the security level. It will be one of the future trends.
What are the technical challenges of face recognition on the access control devices?
There are several technical challenges for face recognition applications, including artificial face discrimination, face recognition, fast response, and low power consumption. In order to rule out the fake conditions, a stereo 3D camera can measure the depth information of the face to distinguish a 2D photo from a 3D face. Furthermore, an IR camera can differentiate a living human face from an artificial model face under IR light illumination. For the face recognition and identification, a new trend is to apply neural network algorithms for better accuracy and robustness. The recognition process can be performed either on the cloud server or at the edge device, depending on the applications. Since the longer the battery life is better for a smart access control device, low power consumption and power saving mode control are very important. Consequently, faster wake up time to accomplish the face recognition task is an essential requirement to ensure good use experience.
How Ci1 intelligent ISP chip solves the facial recognition challenges for access control
iCatch’s Ci1 intelligent image signal processing (ISP) chip has the ability to process both RGB and IR images at the same time while also embeds a neural network acceleration engine for edge computing functions. The ISP engine of Ci1 can connect up to two CMOS sensors to provide stereo image information for further depth processing. Ci1 chip can verify a living face via IR image processing while capturing a color image for actual face recognition. Additionally, the neural processing engine built inside can perform face detection and recognition task locally with better privacy protection and faster response time. Furthermore, Ci1 also has the advantages of low power consumption and fast boot up time which are very much required by the smart access control applications. As a result, Ci1 is also an ideal platform for edge intelligence algorithm development.
The smart access management by face recognition is highly applicable to residential house entrance control with limited members. For example, the room guest can register their face data when checking-in the hotel to link to the designated room number they will stay for entrance control. On the other hand, it can also be used to verify the driver’s authority to a vehicle in fleet control and management. It’s also applicable to identity verification in a payment authentication system.
Ci1 includes an image signal processing engine for both color and IR images, an embedded edge computing acceleration engine, and advantages of low power consumption and fast boot up time. An international technology leader selected iCatch’s Ci1 as the image capturing kernel of its new camera module project and targeted to the intelligent access control system applications. It is expected to enable many potential applications after its mass production.
For more information about iCatch’s products, please visit: www.icatchtek.com
For sales contact, please leave your message at: https://www.icatchtek.com/Requests