The Embedded SDK is for use when the facial recognition software needs to run on a device with limited resources — RAM, CPU, and memory — without any internet or cloud connectivity. However, the device must still perform, for example, detection and recognition of images with varying conditions, such as lighting, or hats and glasses, with a minimum response time.
Using this SDK, developers can create applications that can:
- Provide information on the total number of known/unknown faces, the relative coordinates of each person detected, the size in pixels for each face, and more.
- Learn new faces through training (labeling).
The SAFR Embedded SDK ships as a
tar.gz file that contains the necessary libraries, recognizer, and detector model files, and a sample command line tool called Panoptes.
For SAFR Embedded SDK system requirements, go to SAFR.com.