Systems and methods for analyzing time series data based on event transitions
A method for analyzing time series data to identify an event of interest is provided. The method includes receiving, at a computing device, time series data that includes the event of interest, identifying, using the computing device, a start time of the event of interest and an end time of the event of interest by modeling at least one transitional pattern in the time series data, and categorizing, using the computing device, the event of interest based on the at least one transitional pattern.
1. A method for analyzing time series data to identify an event of interest, said method comprising:
receiving, at a computing device, image time series data that includes the event of interest, the event of interest corresponding to an action performed by a subject;
identifying, using the computing device, a start time of the event of interest and an end time of the event of interest by modeling at least one transitional pattern in the image time series data based on one or more estimated poses of the subject; and categorizing, using the computing device, the event of interest based on the at least one transitional pattern.
2. The method in accordance with claim 1 , wherein the image time series data is video footage.
3. The method in accordance with claim 1 , wherein the event of interest corresponds to an action performed by a single subject.
4. The method in accordance with claim 1 , wherein the event of interest corresponds to an action performed by multiple subjects.
5. The method in accordance with claim 1 , wherein modeling at least one transitional pattern comprises:
modeling an onset segment based on a first of the one or more estimated poses of the subject that represents a transition between a first neutral event and the event of interest; and
modeling an offset segment based on a second of the one or more estimated poses of the subject that represents a transition between the event of interest and a second neutral event.
6. The method in accordance with claim 1 , wherein identifying a start time of the event of interest and an end time of the event of interest comprises identifying the start and end times based on a probability that the event of interest will follow a first neutral event and that a second neutral event will follow the event of interest.
7. The method in accordance with claim 1 , wherein categorizing the event of interest comprises categorizing the event of interest as corresponding to suspicious activity, the method further comprising generating an alert to notify a user that the event of interest corresponds to suspicious activity.
8. A computing device for analyzing time series data to identify an event of interest, the computing device comprising:
a memory device; and
a processor communicatively coupled to said memory device, said processor configured to:
receive image time series data that includes the event of interest, the event of interest corresponding to an action performed by a subject;
identify a start time of the event of interest and an end time of the event of interest by modeling at least one transitional pattern in the image time series data based on one or more estimated poses of the subject; and
categorize the event of interest based on the at least one transitional pattern.
9. The computing device in accordance with claim 8 , wherein to receive image time series data, said processor is configured to receive video footage.
10. The computing device in accordance with claim 8 , wherein to receive image time series data, said processor is configured to receive time series data that includes the action performed by a single subject.
11. The computing device in accordance with claim 8 , wherein to receive image time series data, said processor is configured to receive time series data that includes the action performed by multiple subjects.
12. The computing device in accordance with claim 8 , wherein to model at least one transitional pattern, said processor is configured to:
model an onset segment based on a first of the one or more estimated poses of the subject that represents a transition between a first neutral event and the event of interest; and
model an offset segment based on a second of the one or more estimated poses of the subject that represents a transition between the event of interest and a second neutral event.
13. The computing device in accordance with claim 8 , wherein to identify a start time of the event of interest and an end time of the event of interest, said processor is configured to identify the start and end times based on a probability that the event of interest will follow a first neutral event and a second neutral event will follow the event of interest.
14. The computing device in accordance with claim 8 , wherein said processor is further configured to:
categorize the event of interest as corresponding to suspicious activity; and
generate an alert to notify a user that the event of interest corresponds to suspicious activity.
15. At least one non-transitory computer-readable storage media having computer-executable instructions embodied thereon, wherein when executed by at least one processor, the computer-executable instructions cause the processor to:
receive image time series data that includes the event of interest, the event of interest corresponding to an action performed by a subject;
identify a start time of the event of interest and an end time of the event of interest by modeling at least one transitional pattern in the image time series data based on one or more estimated poses of the subject; and
categorize the event of interest based on the at least one transitional pattern.
16. The at least one non-transitory computer-readable storage media in accordance with claim 15 , wherein to receive image time series data, the computer-executable instructions cause the processor to receive video footage.
17. The at least one non-transitory computer-readable storage media in accordance with claim 15 , wherein to receive image time series data, the computer-executable instructions cause the processor to receive image time series data that includes the action performed by a single subject.
18. The at least one non-transitory computer-readable storage media in accordance with claim 15 , wherein to receive image time series data, the computer-executable instructions cause the processor to receive image time series data that includes the action performed by multiple subjects.
19. The at least one non-transitory computer-readable storage media in accordance with claim 15 , wherein to model at least one transitional pattern, the computer-executable instructions cause the processor to:
model an onset segment based on a first of the one or more estimated poses of the subject that represents a transition between a first neutral event and the event of interest; and
model an offset segment based on a second of the one or more estimated poses of the subject that represents a transition between the event of interest and a second neutral event.
20. The at least one non-transitory computer-readable storage media in accordance with claim 15 , wherein to identify a start time of the event of interest and an end time of the event of interest, the computer-executable instructions cause the processor to identify the start and end times based on a probability that the event of interest will follow a first neutral event and a second neutral event will follow the event of interest.