Unsupervised learning of events in a video sequence

A method of learning events contained within a video image sequence, the method comprising:

• providing a computing system that is configured to receive the image sequence, the computing system programmed to:
• provide a behavioral analysis engine that is configured to learn new events contained within the image sequence, the behavioral analysis engine including an actor feature database, a frame feature database, and a programmable event library stored in a memory;
• initiate a training phase mode within the behavioral analysis engine and obtaining a feature vector including one or more parameters relating to an object disposed within the image sequence;
• analyze the feature vector to determine a number of possible event candidates;
• prompt a user to confirm whether a detected event candidate is a newly identified event;
• and
• store the new event within the event library if the detected event candidate is confirmed by the user.

A method of learning events contained within a video image sequence, the method comprising:

• providing a computing system that is configured to receive the image sequence, the computing system programmed to:
• provide a behavioral analysis engine that is configured to learn new events contained within the image sequence;
• allow a user to initiate a training phase mode within the behavioral analysis engine and obtaining a feature vector including one or more parameters relating to an object disposed within the image sequence;
• apply a time-consistency filtering routine to the image sequence;
• analyze the feature vector to determine a number of possible event candidates;
• prompt a user to confirm whether a detected event candidate is a newly identified event;
• and
• store the newly identified event within an event library if the detected event candidate is confirmed by the user.

A method of learning events contained within a video image sequence, the method comprising:

• providing a computing system that is configured to receive the image sequence, the computing system programmed to:
• provide a behavioral analysis engine that is configured to learn new events contained within the image sequence;
• initiate a testing phase mode within the behavioral analysis engine;
• detect the occurrence of one or more possible event candidates;
• receive confirmation from a user whether a detected event candidate is a newly identified event;
• and
• store the newly identified event within an event library if the detected event candidate is confirmed by the user.

A method of learning events contained within a video image sequence, the method comprising the steps of:

• providing a computing system that is configured to receive the image sequence, the computing system programmed to:
• provide a behavioral analysis engine that is configured to learn new events contained within the image sequence;
• detect the occurrence of one or more events learned by the behavioral analysis engine;
• determine the probability that an event has occurred;
• and
• output a response to a user if the probability that the event has occurred is greater than a confidence threshold value.