1 Overview
Faces can be generally categorized into three types: real faces, faces in images, and faces in videos. Living refers to the real living human faces, while non-living refers to the faces in images or videos.
Liveness detection can distinguish real faces from photos, videos, and other spoofing and automatically determine if the face on the screen is the face of a real person. By accurately detect printed photos and images and videos displayed on the screen, spoofing attacks can be effectively prevented.
1.1 Technical principles
Silent liveness detection technology is divided into three parts: face detection, liveness detection, and face image acquisition.
Face detection, positioning and capture: the system uses algorithm to find the position of the face from the image taken by the camera to determine the position, and the face size is calculated according to its characteristic value, and then process it.
Livneess detection: learn the characteristics and differences of living and non-living bodies through deep learning methods, determine whether the face image is captured from a living body, and effectively prevent attacks by photos, video remakes, and mask models.
Face image acquisition: collect high-quality face images during the detection process.
1.2 Function introduction
Silent liveness detection provides an anti-spoofing detection solution by collect face images through mobile phone cameras and determine whether the face images come from the correct person.
After completing the liveness detection, the returned face image and encrypted binary stream file can be used as input parameters to call the API server (private cloud) interface for face comparison.
1.3 Current version and compatibility
The latest version of the Silent Liveness SDK (iOS version) is V2.4.3 (Updated on June 24, 2021).
The size of the installation package is 9.29MB.
It supports iOS 7.0 or later systems, and iPhone 5 or later models, support two CPU architectures of armv7 and arm64.
1.4 Sample
Run the sample project TestSTSilentLivenessController and click [Start Detection] button to start the liveness detection.
During the liveness detection process, prompts are given according to face status.
After successful detection, corresponding face image will be returned.
1.5 License introduction
Open the license file (xxx.lic) to view the corresponding license expiration period Expiration and license identifier License SN.
If package name error is reported after importing a license into a project, check whether the package name entered in the project is included in the list of bundled package names in the license. If license file is invalid is reported, check whether the imported file is correct and if the problem persists, please contact the support stuff and provide the License SN in the license file.
1.5.1 How to use
The provided SDK sample does not contain the license file by default. After obtaining the license file, the file must be renamed and placed into the specified path before running. The authorization method of SDK sample for iOS silent liveness detection is described in detail as follows:
1.5.2 Trial license instructions
Rename the license file (xxx.lic) to SenseID_Liveness_Silent.lic
Place the SenseID_Liveness_Silent.lic file in the libSTSilentLivenessController folder
Check whether the SenseID_Liveness_Interactive.lic file exists in Project-->TARGETS-->Build Phases-->Copy Bundle Resources
Ensure that the name of the license file is the same as the one in the project code
1.5.3 Official license instructions (requires package name)
Rename the license file (xxx.lic) to SenseID_Liveness_Silent.lic
Place the SenseID_Liveness_Silent.lic file in the libSTSilentLivenessController folder
Check whether the SenseID_Liveness_Interactive.lic file exists in Project-->TARGETS-->Build Phases-->Copy Bundle Resources
Ensure that the name of the license file is the same as the one in the project code
View project-> TARGETS-> General-> identity->, and modify the package name corresponding to Bundle Idenfifier into the package name that is actually bound to the license.
Last updated
