IDmission > Facial Recognation
Passive Liveness for Facial Recognation
IDmission is a leader in the use of Passive Liveness for identity verification. With over a decade of development of Machine Learning based facial recognition apps, IDmission’s passive liveness is used worldwide to eliminate identity fraud and use facial biometrics to build identity databases.
Passive Liveness Detection
Our passive liveness detection automatically detects and rejects presentation attacks, including photos and videos, print-outs, masks and deep fakes. IDmission is one of the few companies meeting global ISO 30107-3 Presentation Attack Detection (PAD) Level 2 standards.
Machine Learning Liveness Models
IDmission Machine Learning models are trained on millions of identity records from over 200 countries, creating a globally effective and unbiased platform for building identity systems.IDmission uses a convolutional neural network (CNN) to solve the problem of detecting liveness from realtime images.
The CNN uses deep learning (multiple layers) as follows:
- Layers - 160 layers deep
- Trainable Parameters: 3,422,000
- Performs a 3D convolution using 3 pixel x 3 pixel x 1 pixel filters
Each day, tens of thousands of new facial verifications are performed, with results from all IDmission customers continuously used to sharpen the models.
Capture on web page and
verify against server
Liveness checks with single
frame or multiple frames
Passive Liveness Facial Recognition Highlights
Machine learning technology
Trained extensively to detect spoofing attacks that are invisible to the naked eyes
Invisible by design
Completely passive. No need to blink, smile, turn head, or zoom in/out
Performance based on real world data
100,000 faces from Latin America, Africa, North America, South Asia and Asia Pacific