METHOD FOR FINDING PATHS IN VIDEO
A system for detecting behavior of a target may include: a target detection engine, adapted to detect at least one target from one or more objects from a video surveillance system recording a scene; a path builder, adapted to create at least one mature path model from analysis of the behavior of a plurality of targets in the scene, wherein the at least one mature path model includes a model of expected target behavior with respect to the at least one path model; and a target behavior analyzer, adapted to analyze and identify target behavior with respect to the at least one mature path model. The system may further include an alert generator, adapted to generate an alert based on the identified behavior.
1 . A system for detecting behavior of a target, comprising:
a target detection engine, adapted to detect at least one target from one or more objects from a video of a scene;
a path builder, adapted to create at least one mature path model from analysis of the behavior of a plurality of targets in the scene, wherein said at least one mature path model includes a model of expected target behavior with respect to said at least one path model; and
a target behavior analyzer, adapted to analyze and identify target behavior with respect to said at least one mature path model.
2 . The system according to claim 1 , wherein said target behavior analyzer is adapted to identify behavior inconsistent with the model of expected behavior, comprising detecting at least one of:
a target deviating from the path;
a target traveling off of the path;
a target switching from the path to another path;
a target crossing the path;
a target traveling the path, wherein the path is infrequently traveled;
a target traveling at an unusual speed on the path;
a target stopping at an unusual stopping point on the path;
a target traveling at an unusual time on the path;
a target traveling in an unusual direction on the path;
a target of a type not normally found on the path; or
a target having an unusual physical property on the path.
3 . The system according to claim 1 , wherein said target behavior analyzer is adapted to predict a target's subsequent path based on said at least one path model and the target's observed behavior.
4 . The system according to claim 1 , wherein said target behavior analyzer is adapted to classify at least one of the target type, based on a path type, or the path type based on the target type.
5 . The system according to claim 1 , wherein the target behavior analyzer is further adapted to build a statistical model of target behavior with respect to the path model.
6 . The system according to claim 5 , wherein the target behavior is analyzed with respect to a statistical model of at least one of:
a target dwell time duration on the path;
a target dwell location on the path;
a target property associated with at least one of: the path, a dwell location on the path, and/or a dwell time on the path;
an interaction of a target with another target of the same type on the path;
an interaction of a target with another target of a different type on the path;
an interaction of a target with an element in the scene;
a temporal pattern of a target property on the path; or
a deviation from normal target properties on the path.
7 . The system according to claim 6 , wherein the target behavior analyzer is adapted to identify target behavior from a combination of:
at least two detected behaviors inconsistent with a path model;
at least two target detected behaviors with respect to a statistical model; or
at least one detected inconsistent behavior and at least one detected behavior with respect to a statistical model.
8 . The system according to claim 1 , wherein said video is received from a video surveillance system.
9 . The system according to claim 1 , wherein said system is implemented in application-specific hardware to emulate at least one of a computer or software.
10 . The system according to claim 1 , further comprising an alert generator, adapted to generate an alert based on said identified behavior.
11 . A computer-based method of target behavior analysis, comprising:
processing an input video sequence to obtain target information for at least one target from one or more objects from a video of a scene;
building at least one mature path model from analysis of the behavior of a plurality of targets in the scene, wherein said at least one mature path model includes a model of expected target behavior with respect to said at least one path model; and
analyzing and identifying target behavior of a target with respect to said at least one mature path model.
12 . The method according to claim 11 , further comprising:
generating an alert based on said identified target behavior.
13 . A computer-readable medium containing instructions that, when executed by a processor, cause the processor to perform the method according to claim 11 .
14 . A video processing system comprising:
a computer system; and
the computer-readable medium according to claim 13 .
15 . A video surveillance system comprising:
at least one camera to generate an input video sequence; and
the video processing system according to claim 14 .
16 . The method according to claim 11 , wherein processing an input video sequence comprises processing video from a video surveillance system.
17 . The method according to claim 11 , wherein said analyzing and identifying target behavior includes identifying target behavior inconsistent with the model of expected behavior, comprising detecting at least one of:
a target deviating from the path;
a target traveling off of the path;
a target switching from the path to another path;
a target crossing the path;
a target traveling the path, wherein the path is infrequently traveled;
a target traveling at an unusual speed on the path;
a target stopping at an unusual stopping point on the path;
a target traveling at an unusual time on the path;
a target traveling in an unusual direction on the path;
a target of a type not normally found on the path; or
a target having an unusual physical property on the path.
18 . The method according to claim 11 , further comprising predicting a target's subsequent path based on said at least one path model and the target's observed behavior.
19 . The method according to claim 11 , wherein said analyzing and identifying target behavior includes classifying at least one of the target type, based on a path type, and/or the path type based on the target type.
20 . The method according to claim 11 , wherein said analyzing and identifying target behavior further includes building a statistical model of target behavior with respect to the path model.
21 . The method according to claim 20 , wherein said analyzing and identifying target behavior includes analyzing target behavior with respect to a statistical model of at least one of:
a target dwell time duration on the path;
a target dwell location on the path;
a target property associated with at least one of: the path, a dwell location on the path, and/or a dwell time on the path;
an interaction of a target with another target of the same type on the path;
an interaction of a target with another target of a different type on the path;
an interaction of a target with an element in the scene;
a temporal pattern of a target property on the path; or a deviation from normal target properties on the path.
22 . The method according to claim 21 , wherein said analyzing and identifying target behavior includes identify target behavior from a combination of:
at least two detected behaviors inconsistent with a path model;
at least two target detected behaviors with respect to a statistical model; or
at least one detected inconsistent behavior and at least one detected behavior with respect to a statistical model.
23 . A computer-readable medium containing instructions that, when executed by a processor, cause the processor to perform operations comprising:
processing an input video sequence to obtain target information for at least one target from one or more objects from a video of a scene;
building at least one mature path model from analysis of the behavior of a plurality of targets in the scene, wherein said at least one mature path model includes a model of expected target behavior with respect to said at least one path model; and
analyzing and identifying target behavior of a target with respect to said at least one mature path model.
24 . The computer-readable medium according to claim 23 , the operations further comprising:
generating an alert based on said identified target behavior.
25 . The computer-readable medium according to claim 23 , wherein processing an input video sequence comprises processing video from a video surveillance system.
26 . The computer-readable medium according to claim 23 , wherein said analyzing and identifying target behavior includes identifying target behavior inconsistent with the model of expected behavior, comprising detecting at least one of:
a target deviating from the path;
a target traveling off of the path;
a target switching from the path to another path;
a target crossing the path;
a target traveling the path, wherein the path is infrequently traveled;
a target traveling at an unusual speed on the path;
a target stopping at an unusual stopping point on the path;
a target traveling at an unusual time on the path;
a target traveling in an unusual direction on the path;
a target of a type not normally found on the path; or a target having an unusual physical property on the path.
27 . The computer-readable medium according to claim 23 , the operations further comprising predicting a target's subsequent path based on said at least one path model and the target's observed behavior.
28 . The computer-readable medium according to claim 23 , wherein said analyzing and identifying target behavior includes classifying at least one of the target type, based on a path type, or the path type based on the target type.
29 . The computer-readable medium according to claim 23 , wherein said analyzing and identifying target behavior further includes building a statistical model of target behavior with respect to the path model.
30 . The computer-readable medium according to claim 29 , wherein said analyzing and identifying target behavior includes analyzing target behavior with respect to a statistical model of at least one of:
a target dwell time duration on the path;
a target dwell location on the path;
a target property associated with at least one of: the path, a dwell location on the path, and/or a dwell time on the path;
an interaction of a target with another target of the same type on the path;
an interaction of a target with another target of a different type on the path;
an interaction of a target with an element in the scene;
a temporal pattern of a target property on the path; or a deviation from normal target properties on the path.
31 . The computer-readable medium according to claim 30 , wherein said analyzing and identifying target behavior includes identify target behavior from a combination of:
at least two detected behaviors inconsistent with a path model;
at least two target detected behaviors with respect to a statistical model; or
at least one detected inconsistent behavior and at least one detected behavior with respect to a statistical model.