In the RESEARCH module, people can be tracked across cameras (reidentification/Re-ID) using the face recognition functionality when a high-quality face recognition camera is used. This feature can be used to count the number of distinct visitors according to face in store, measure the average time people stay in a store, calculate bounce rates, exclude employees from visitor counts and measure new versus repeat visitors.

This feature can be used for people whose faces are visible and of high quality (2-star and 3-star images) on all the relevant cameras.

To support this feature, RESEARCH has many built-in measures, such as # of Unique Identities, # of New Unique Identities, # of Returning Identities, Avg Visit Duration Unique Identities, # of Bounced Visitors, Visitors Bounce Rate, and more. In order to use this feature, all relevant camera sources need to be added to the same Source Group in the SOURCE GROUPS tab.

The Re-ID mechanism can be modified to account for the general similarity between people in the environment. For example, if most people wear a hair covering or have distinct facial features, such as mustaches or glasses, it may be necessary to adjust the sensitivity of this feature using the BIFaceRecognition.FaceMatchingThreshold setting in the BriefCam Administrator Console.