Method and system for operating a video surveillance system including predicting and displaying possible future activity of a person of interest

A video surveillance system of a facility, comprising:

• a plurality of video cameras distributed about the facility providing a plurality of video streams;
• a video recorder recording the plurality of video streams;
• an operator console;
• a controller operatively coupled to the plurality of video cameras, the video recorder and the operator console, the controller configured to:
• process each of the video streams using video analytics to identify people in the plurality of video streams;
• identify an intent of each of the identified people based at least in part on one or more of the plurality of video streams;
• classify the identified intent of each of the identified people into one of a plurality of intent categories, wherein one of the plurality of intent categories corresponds to an intent of concern;
• identify a person of interest as a person that has an identified intent that is classified as an intent of concern, and for the person of interest:
• use one or more trained Artificial Intelligence (AI) models to predict a future activity of the person of interest based at least in part on the identified intent of the person of interest;
• provide a description of the predicted future activity of the person of interest to a Generative AI model, wherein the Generative AI model generates a simulated video stream of the person of interest performing the predicted future activity;
• and
• display the simulated video stream of the person of interest performing the predicted future activity in a prediction salvo on the operator console.

A non-transitory computer readable medium storing instructions thereon that when executed by one or more processors causes the one or more processors to:

• process each of a plurality of video streams using video analytics to identify people in the plurality of video streams;
• use one or more trained Artificial Intelligence (AI) models to identify an intent of each of the identified people based at least in part on one or more of the plurality of video streams;
• classify the identified intent of each of the identified people into one of a plurality of intent categories, wherein one of the plurality of intent categories corresponds to an intent of concern;
• identifying a person of interest as a person that has an identified intent that is classified as an intent of concern, and for the person of interest:
• use one or more trained Artificial Intelligence (AI) models to predict a future activity of a person of interest;
• automatically generate a text description of the predicted future activity of the person of interest;
• automatically provide the text description of the predicted future activity of the person of interest and one or more images of the person of interest extracted from one or more of the video streams to a Generative AI model, wherein the Generative AI model generates a simulated video stream of the person of interest performing the predicted future activity;
• and
• display the simulated video stream of the person of interest performing the predicted future activity in a prediction salvo on an operator console of a video surveillance system.