Home > News > Events

MIPS AI SDK For Face Recognition

Jun. 29, 2019

As the core driving force of the new round of industrial transformation, artificial intelligence has spawned a large number of new technologies and new products, and has become a key player in promoting the innovation and development of Commercial display and security industry, and has also become the commanding height of industrial competition. Therefore, in 2018, SMDT officially entered the “display + AI” and launched the “Face ID Strategy”. With the excellent hardware quality and powerful advantages of the face recognition algorithm, a large number of face recognition projects have been promoted in many fields.

In the process of promoting the Face Recognition Technology, it seems that SMDT found that the market demand for face recognition is the strongest in the field of commercial display and security, and there are relatively large differences in the requirements of face recognition technology in these two fields. According to SMDT, the MIPS AI SDK has been optimized and upgraded to better meet the needs of the industry, more professional and more powerful!

Face Recognition Technology

MIPS face recognition API commercial version

1. Overview:

This module provides a basic module for Android applications, which is suitable for application scenarios such as face recognition based on passenger flow statistics, precision promotion, VIP recognition, and powder sucking games. The purpose is to read the camera screen and detect the face on the screen. Identify the age, gender, mood and other characteristics of the face.

2. Typical scene:

Accurate promotion based on face recognition

Game interaction based on face recognition

Face Recognition Technology

MIPS face recognition API version

1. Overview:

This module provides a basic module for the android application, which is suitable for application scenarios such as human witness comparison devices and face recognition gates. The purpose is to read the camera's screen and detect the face on the screen. At the same time, the following operations can be performed:

1) Live detection: Detect whether the face appearing in the picture is a real person or a photo or video spoof. The SDK supports single-head camera and binocular camera (RGB+IR, including live detection) while performing the following operations:

a. Monocular in vivo detection: To detect whether the face appearing in the picture is a real person or a photo or video counterfeit, an RGB (visible light) monocular camera is required. Description: Supports multiple people in the same screen to perform in-vivo detection (selecting faces specified by the application). Only one face on the same screen can be used for monocular detection;

b. Binocular in vivo detection: To detect whether the face appearing in the picture is a real person or a photo or video counterfeit, use an RGB (visible light) + IR (infrared) binocular camera. Description: Support multiple people in the same screen to perform live detection (select face specified by the application);

Face Recognition Technology

2) Face database search (1: N comparison): First, a face database (within 20,000 people) is created by the photo of the VIP person, and the face feature data can be stored in the face database. Then in the real-time detection process: by extracting the features of the face in the camera picture, comparing with the features in the face database (generating the highest similarity), and identifying whether the current face is in the face database according to the similarity. (To ensure the search and comparison accuracy, the face entered into the face database needs to ensure a positive face, and the face frame width is greater than 90 pixels.)

3) Face comparison (1:1 comparison): It is mainly used for similar scenes such as human witness comparison. Compare the photo and camera picture (or photo) to the same person: extract the face feature in the photo and the camera picture separately, then compare the two face features to produce similarity. If the similarity is greater than the set threshold, you can think It is the same person.

2. Typical scene

Face recognition gate

Face Recognition Technology


Inquiry

For all inquiries, please fill in the form below (* are required) to send us a brief message, and we will get back to you as soon as possible.

* Name:
Mobile:
* E-mail:
* Area:
Application:
  • MIPS Content Management Software
  • Face Detection Technology
  • Smart Vending Solution
  • Functional characteristics
  • Pharmaceutical industry
* Message:
*
WELCOME TO SMDT
contact us