Palmprint recognition algorithm| Infrared face recognition algorithm| Fingerprint recognition algorithm| Human-witness comparison algorithm
 

Main functions: A self-developed face recognition algorithm is adopted to achieve fast, convenient and secure self-verification of personal identity information and biometric information. At the same time, collect photos of on-site personnel through high-definition cameras, judge the similarity between the ID card chip photos read and the on-site photos of the holder, confirm whether they are themselves and record the identification results, and collect the corresponding ID card and photo information and color real-time monitoring photo information. The new second-generation natural light face recognition algorithm based on neural network deep learning has greatly improved the face recognition effect, solved the problems of illumination, posture, glasses, and small area occlusion, and has the face feature self-learning function, which can effectively solve the problem of slow change of face over time. Deep learning can be carried out based on faces with different illumination, viewpoint, age, identity and expression, and the face information successfully verified on the platform can be automatically established in the database as a template for subsequent verification. Support Windows, Linux, Android and IOS and other platforms, face detection, face recognition, gender, age, expression recognition, distance detection, posture detection, living detection, face synthesis and deformation, a variety of animal recognition, technology is constantly updated to adapt to more complex environments and applications. The recognition rate is as high as 99%, and the pass rate is 98.3%.

Comparison process: intelligent face detection: accurately locate the face position and size in the image, and detect the effective image; Intelligent verification (human-witness comparison) : The feature data of the facial image is extracted and compared with the feature data in the ID card photo, and the matching result is output through the threshold setting. Fingerprint verification module is optional, multiple verification, more secure; Application: Used in hotel check-in, Internet cafe registration, security check, logistics, education, visitor management, community access management, telecommunications real-name management, financial real-name management, government agencies real-name management and other major industries. The algorithm is cloud-based comparison and front-end embedded offline comparison, applicable to various platform docking and terminal equipment.

Specification parameter

Human-witness comparison Performance parameters (related to hardware configuration) :

Face recognition core PC: simplified version of the program itself size 25M, advanced version 65M; Operation space <100M;

Face feature template size: single <1.5KB;

Supported platforms: Android, IOS, Windows, Linux and embedded Linux;

Hardware requirements: processor frequency > 1GHz;

Minimum hardware requirements for android core: armeabi-v7a, memory ≥ 1G, android 4.0 or above;

Face recognition:

The ideal distance between eyes is more than 40 pixels, and the ideal distance between eyes is more than 50 pixels.

Can recognize the face image, the correct recognition rate of more than 98%

Support multi-face positioning, a single picture can be positioned to identify more than 8 people;

Can identify the age of people, children, adolescents, youth, middle-aged, elderly five stages, age recognition accuracy of more than 80%; Can identify the gender of the person, identification accuracy of more than 90%; They can recognize five expressions: happy, angry, surprised, sad, and cool. Image specifications: the distance between the eyes and the face of the photo is more than 40 pixels; Camera resolution: not less than 300,000 pixels; Head Angle: positioning up and down around plus or minus 45 degrees, identify up and down around plus or minus 20 degrees;

Technology licensing methods:

At present, core authorization is our main authorization method, which provides users with the recognition core dynamic library, and customers can realize face detection and face recognition according to our interface

Supported platform: Android (java), Windows (c++/c#/java);

File provided: image.png

Encryption method:

PC (windows) : Dongle encryption/network authentication;

Android: Network authentication/encryption chip;

Service hotline

+86-0755-2607 5207

SHENZHEN FUGE TECHNOLOGIES CO.,lTD.

Add.:Rom 205,Building A1,Jianyuan Zhichuangyuan,Xixiang Street,Bao'an District,Shenzhen

Fax:+86-0755-26075207

Email:lily@fugesys.com

Copyright © 2016-20182017 深圳福鸽科技有限公司 All Rights Reserved.
ICP备:粤ICP备17084424号