Using Machine Learning to Focus on an Object of Interest During Remote Collaboration

As the need for world-wide, on-demand, real-time technical support increases, remote collaboration solutions are enabling critical virtual telepresence capabilities for remote Subject Matter Experts (SME). Whether the use case is providing remote technical support during assembly, test, inspection, integration, or troubleshooting and repair, today’s commercial remote collaboration solutions face practical utilization challenges. Remote collaboration solutions rely on two-way video streaming. The deployment of these solutions is often restricted, or severely limited, to specific use cases in specific locations because of potential visual data spills. Mitigation of this risk currently means limiting where cameras are allowed and moving objects of interest (OOI) into rooms where cameras are deemed safe to use. Such limitations create a need for automatic filtering of live video streams which is easy for manufacturing and maintenance workers to use.