Unsupervised learning of events in a video sequence
First Claim
1. 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 training phase mode within the behavioral analysis engine and obtain a feature vector including one or more parameters relating to an object disposed within the image sequence;
identify one or more clusters for at least some of the one or more parameters, at least some of the one or more clusters corresponding to possible event candidates;
display an identifier for at least some of the possible event candidates on a display; and
allow a user to select one or more of the possible event candidates, and to include the selected one or more of the possible event candidates into an event library.
1 Assignment
0 Petitions
Accused Products
Abstract
Methods and systems for the unsupervised learning of events contained within a video sequence, including apparatus and interfaces for implementing such systems and methods, are disclosed. An illustrative method in accordance with an exemplary embodiment of the present invention may include the steps of providing a behavioral analysis engine, initiating a training phase mode within the behavioral analysis engine and obtaining a feature vector including one or more parameters relating to an object located within an image sequence, and then analyzing the feature vector to determine a number of possible event candidates. The behavioral analysis engine can be configured to prompt the user to confirm whether an event candidate is a new event, an existing event, or an outlier. Once trained, a testing/operational phase mode of the behavioral analysis engine can be further implemented to detect the occurrence of one or more learned events, if desired.
-
Citations
40 Claims
-
1. 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 training phase mode within the behavioral analysis engine and obtain a feature vector including one or more parameters relating to an object disposed within the image sequence; identify one or more clusters for at least some of the one or more parameters, at least some of the one or more clusters corresponding to possible event candidates; display an identifier for at least some of the possible event candidates on a display; and allow a user to select one or more of the possible event candidates, and to include the selected one or more of the possible event candidates into an event library. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19)
-
-
20. 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. - View Dependent Claims (21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35)
-
-
36. 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.
-
-
37. 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. - View Dependent Claims (38, 39)
-
-
40. 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.
-
Specification