The IsSceneChange property signals that the object was tracked across a key frame, which means the object may have been tracked across a scene change. Thus,...

Background SAFR can be used to measure attention span or gaze duration of people in the camera view.  It does this by reporting when people are looking dir...

Some platforms have dependencies on additional 3rd party software.  See the System Requirements topic for a complete list of requirements for each platform.

Following describes how to create a video frame from an image and process it with the ObjectTracker.  See https://docs.real.com/sdk/windows/Windows_SDK.html...

The same REST APIs are supported in cloud and local deployments: the SAFR Web APIs described near the bottom of the developers page of the SAFR Download Por...

In order to perform face, person or liveness detection, SAFR must generate model files.  SAFR does this as a one time action upon first run.  On subsequent ...

The benchmarks below are for an A72 Cortex system on a chip (SOC). Facedetector depends on the input image but it usually requires around 200ms.   Eac...

This article explains how to add more than one face image to an existing identity. Method 1 - Merging Identities To add a new face image to existing id...

This document will describe how to use the REST APIs to queue two or more files to be processed in sequence.  This process requires use of SAFR REST APIs. ...

rootPersonId is added by Identity Sync and refers to the identity on origin server. mergeWith is used to refer to parent of merged face identities. merg...