Liveness Detection |How To Choose Suitable Camera Of Face Recognition Terminal



The safety of face recognition technology gradually gains attention as face recognition is widely applied at various scenes. Particular in unattended application, liveness detection is the indispensable component.
Nowadays, there are two main categories of liveness detection, one is active, and another is passive. In the market, common face recognition terminals will introduce monocular/binocular /3D structured light cameras to support liveness detection. However, what are the differences and scene applications between monocular, binocular and 3D structured light cameras? Today, let's take a look at it. 
Monocular camera (RGB) parsed unstructured video data to structure data through monocular algorithm in real time. Based on people's facial features, it can judge whether there is face exist in the input face images or video and capture face key information and analysis gender, age and other attributes.
It can be used for accurate marketing, facial interaction or customer sentiment analysis. The monocular RGB liveness detection supports 1: N face database search and 1:1 witness comparison.
Advantage: easy to deploy and low cost
Scene: accurate marketing, passenger flow statistic, access control management, etc
Applicable equipment: face recognition attendance terminal, self-service device, smart shelf
Binocular camera (RGB + IR) can capture and recognize face, make accurate identification at some key feature points include eyes, ears and nose , and calculate space information such as pupil distance, nose height, distance from eyes to mouth to ears and so on within 1mm error.
Advantage: Higher accuracy. It can effectively resist attacks from photos and videos, and has stronger adaptability to light change, complex background environment and other factors. 
Scene: Access control management
Applicable equipment: face recognition gate, face recognition attendance machine, etc.
3D structured light. Based on deep learning structured light algorithm and 3D depth sensing camera, it can use structured light device to obtain color, infrared, depth photo of scene. Through detect and analysis face, it will form 3D face image and create facial model with facial depth information, achieve better recognition speed, precision and security.
Advantage: Superior security, recognition accuracy and recognition speed. It can more effective against attacks from paper, mask and other tool and not effected by ambient light. 
Scene: face recognition payment, places with higher safety demand
Applicable device: face recognition payment, smart vending machine
In conclusion, the cost of monocular camera is lower, but it should combine with algorithm to increase accuracy. Binocular camera with reasonable price and better accuracy is most widely used in the market. 3D structured light owns the best performance and the expensive price, more suitable for abundant budget places and projects.


评论

此博客中的热门博文

How To Clean And Maintain Self-checkout Kiosk?

Health Pass Solution Assist Vaccination Status Verification

A Comprehensive Guide on NFC POS Machines