Video framing based on tracked characteristics of meeting participants

Abstract

Multi- and single-camera video conferencing systems are disclosed including at least one camera, each configured to generate an overview video output stream representative of a meeting environment, and at least one video processing unit. The video processing unit(s) are configured to automatically analyze the overview video output stream(s) to identify one or more meeting participants and assign respective role designations to the meeting participants based on analysis of the overview video output stream(s). The video processing unit(s) are further configured to generate, and output for display, one or more primary video streams including one or more of the meeting participants, based on the assigned roles.