Skip to content

Latest commit

 

History

History

FaceIdentification

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 

SeetaFace2 Identification

License

Description

SeetaFace2 Identification is designed with Deep Residual Network, which has shown impressive performance for image classification. With the input face size of 256x256, the network inference takes about 100ms for feature extraction per face on I7-CPU, which is a trade-off between accuracy and time cost. The released model is trained with 33 million face images of more than 320K subjects.

How to run SeetaFace2 Identification (C++)

To detect faces, one should first instantiate an object of seeta::FaceRecognizer2 and initialize it with path of the model file. Then one can call Compare to compare faces. See more details at under links.

Notice

SeetaFace2 Identification only support 1 instance in single application.

SeetaFace2 Identification only support 500 probes in memory.

Documents

Downloads

License

SeetaFace2 Identification is released under the BSD 2-Clause license.