Face recognition privatization deployment package_supporting mega face database Baidu AI open platform

The face privatization model can provide face detection, face comparison, face search, live detection, face recognition with masks and other functions. The identification accuracy and response speed of each functional interface are leading the industry, and have won international awards for many times. The model package supports mainstream CPU/GPU environment and localization system deployment, which can be deployed to local physical machines, local virtual machines and virtual machine servers with one click to ensure data privacy and security.

Function description

Program composition

Application scenarios

Characteristic advantages

Product pricing

Related recommendations

Function description

Face Detection and Attribute Analysis

Accurately locate 72 key points of face, quickly identify a variety of facial attribute information, including age, gender, mood, whether to wear glasses, etc; It can control the quality of face images to ensure that the images meet the quality standards and ensure the accuracy of subsequent face comparison and face search.

Face contrast

It supports face comparison of 5 image types, including life photos, ID card photos, ID card chip photos, textured photos, infrared black and white photos. It can compare the similarity of two faces in the picture and return the similarity score.

Face database management

It supports ten million level face database management, and provides 15 interface capabilities such as face registration, update, deletion, user information query, and full level face visualization management interface. Users can realize base database migration through face database management function.

Face Search

It can realize millisecond level recognition response, support the comparison of N faces in the face database, can carry out 1: N retrieval or M: N retrieval, and return the similarity score. It can meet the application scenarios of face attendance, face brushing, security monitoring, etc.

In vivo detection

Provide rich live detection services, including live detection of pictures, live detection of video actions, live detection of lip language, and live function of dazzle pupil. The live detection also includes the composite image capability to identify whether the fake face data, such as face changing, PS, face fusion, etc., are tampered with by the late synthetic video, effectively resist cheating and ensure business security.

Face recognition with masks

It can perform face recognition for users wearing masks. By locating users' facial key points, accurately analyzing their facial feature attributes, and performing comparison and search services, it can meet medical and epidemic prevention and other application scenarios.

Programme composition

Identity verification scheme

Face Attendance Scheme

Face brushing access scheme

Video monitoring scheme

Identity verification scheme

The scheme consists of face privatization, OCR ID card recognition, face acquisition SDK, and security risk control products, providing customers with active and passive identity verification schemes with high security guarantees.

Face privatization

The face privatization model can provide face detection, face comparison, face search, live detection and other functions

Face Acquisition SDK

Offline call face detection, face tracking, face collection and other capabilities to quickly obtain high-quality face photos View details

OCR ID card identification

Structured identification of all 8 fields on the front and back of the second-generation resident ID card, with an accuracy rate of more than 99% View details

Application scenarios

Identity verification

Face Attendance

Face brushing access

Video surveillance

Identity verification

The identity verification scheme is mainly applied to face login, face payment and special identity recognition scenarios. Among them, the face login scenarios include but are not limited to the mobile phone face unlock screen, payment software face login, etc. Face swiping payment scenarios include but are not limited to remote banking business transactions, payment software face swiping transactions, etc. Special identity recognition scenarios include but are not limited to online taxi hailing driver/passenger identity recognition, member identity recognition, minor identity recognition under social game protection mode, etc. This scheme can effectively manage user information and verify the authenticity of target users in time. At the same time, based on the secure encrypted transmission mode and the privatization deployment mode, the user information security can be guaranteed during verification to prevent the disclosure of private data.

Cooperation cases

Characteristic advantages

Data security and privacy

The face privatization model can be deployed to a local server, and data storage and processing are carried out in the enterprise intranet, without uploading to a third party, to ensure the privacy and security of the enterprise's core production data.

Millisecond response

The face privatization model supports the management and retrieval of tens of millions of databases, and can realize millisecond response and recognition. It can achieve high throughput and low latency under high concurrency to ensure normal business operation.

Leading recognition effect

The recognition effect of each functional interface in the face privatization model is leading in the industry, and has won the authoritative certification of the industry, and won international awards for many times. Among them, the mask recognition function won the first place in CSIG FAT-AI open scene mask face recognition.

Professional after-sales service

The face privatization team can provide complete after-sales service support to ensure that users can respond to their needs in 7 * 24 hours. In case of technical accidents, the emergency response team will investigate and solve them in time to reduce the impact of accidents and ensure the normal operation of customers' main business.

Localized adaptation

The face privatization model can run in the mainstream CPU/GPU environment and the localization system. Among them, the CPU supports the version with performance above 2.3 GHz, the GPU supports P4/T4/RTX system and domestic video card, and the concurrent performance of a single card can reach up to 80QPS.

Convenient and flexible deployment

The face privatization model supports one click deployment to local physical machines, local virtual machines, and virtual machine servers, and the deployment service can be completed in half an hour at the earliest. Update authentication services can be provided in a timely manner according to business needs.

Product pricing

Price of pure software version

Price of hardware and software all-in-one machine

Normal Edition

Basic functions of face recognition

Remote Assistance Deployment

First year free maintenance

Authorized price

Business inquiry

Apply Now

Advanced Edition

Basic functions of face recognition