Method and system for detecting semantic events

• US 7,324,984 B2
• Filed: 11/14/2003
• Issued: 01/29/2008
• Est. Priority Date: 01/23/2001
• Status: Expired due to Fees

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Tovinkere, Vasanth R., Epshteyn, Eugene, Qian, Richard
• Primary Examiner(s): STARKS, WILBERT L
• Application Number:
  US10/714,124
• Publication Number:
  US 20040102921A1
• Time in Patent Office: 1,537 Days
• Field of Search: 706/55, 706/28, 706/20, 706/45, 706/15, 704/254
• US Class Current: 706/55
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

A method and system is provided for detecting occurrences of semantic temporal events based on observations extracted from input data and event models. The input data is fed into the system from some data source. Based on specified event to be detected, multiple-layer models corresponding to the event are retrieved. The models are used to determine the types of temporal observations to be extracted from the input data. The extracted temporal observations are then used, in combination with the multiple-layer models of the event, to detect the occurrences of the event.

Claims

1. A system for detecting a semantic temporal event included in unprocessed video or audio data of an event from at least one data source, said system comprising:

a knowledge-based modeling unit for generating multiple-layer models for identifying said semantic temporal event;

a storage mechanism for storing said multiple-layer models;

an observation collection unit for extracting, from said unprocessed video or audio data from the at least one data source, temporal observations according to said multiple-layer models for the semantic temporal event; and

a temporal event detection unit for detecting one or more occurrences of the semantic temporal event based on said temporal observations and said multiple-layer models, wherein the semantic temporal event occurs during the event, the event has a start time and an end time, and the detected one or more occurrences of the semantic temporal event are used to facilitate selective delivery of the unprocessed video or audio data.

2. The system according to claim 1, further including:

an event characterization unit for characterizing said one or more occurrences of the semantic temporal event, detected by said temporal event detection unit, to produce a characterization for the one or more occurrences of the semantic temporal event.

3. The system according to claim 2, further including:

a storage mechanism for storing the characterization produced by said event characterization unit;

an event prediction unit for performing temporal event prediction based on said characterization;

an event model updating unit for modifying said multiple-layer models based on said characterization; and

an event simulation unit for simulating parts of said semantic temporal event according to said characterization.

4. The system according to claim 2, further including:

an event storage, in which detected occurrences of temporal semantic events are stored;

an events statistics extractor to compute statistical information about the detected occurrences; and

an event statistics storage unit to store the statistical information.

5. The system according to claim 1, wherein the event is a sports event.

6. The system according to claim 5, wherein said sports event includes a soccer game.

7. The system according to claim 1, wherein said multiple-layer models include a high level domain-specific knowledge model and a dynamic hierarchical event model.

8. The system according to claim 7, wherein said high level domain-specific knowledge model includes rules of a sports game.

9. The system according to claim 7, wherein said dynamic hierarchical event model includes a hierarchical decision tree.

10. The system according to claim 7, wherein said dynamic hierarchical event model includes an entity-relationship-diagram.

11. The system according to claim 1, wherein said at least one data source is selected from the group consisting of a camera, a microwave sensor, a sound recorder, and an input data stream.

12. The system according to claim 1, wherein said at least one data source includes a data stream sent through a network connection.

13. The system according to claim 12 wherein the data stream is a video stream with synchronized audio track.

14. The system according to claim 1, wherein the observation collection unit is simultaneously connected to more than one data source.

15. The system according to claim 1, wherein said temporal event detection unit includes an integration unit, a detection unit, and a fusion unit, and said integration unit combines a plurality of observation streams from a plurality of data sources, the detection unit detects a same event using a plurality of detection means to produce a plurality of detection results, and said fusion unit fuses the plurality of detection results to produce a single detection decision.

16. A system for detecting a semantic temporal event included in unprocessed video or audio data of an event from at least one data source, said system comprising:

a knowledge-based modeling unit for generating multiple-layer models for said semantic temporal event;

a storage mechanism for storing said multiple-layer models;

an observation collection unit for extracting, from said unprocessed video or audio data from said at least one data source, temporal observations according to said multiple-layer models for the semantic temporal event;

a temporal event detection unit for detecting one or more occurrences of the semantic temporal event included in the unprocessed video or audio data based on said temporal observations and said multiple-layer models;

an event characterization unit for characterizing said one or more occurrences of the semantic temporal event, detected by said temporal event detection unit, to produce a characterization for the one or more occurrences of the semantic temporal event;

a storage mechanism for storing the characterization produced by said event characterization unit;

an event prediction unit for performing temporal event prediction based on said characterization;

an event model updating unit for modifying said multiple-layer models based on said characterization; and

an event simulation unit for simulating parts of said semantic temporal event according to said characterization, wherein the semantic temporal event occurs during the event and the event has a start time and an end time.

17. The system according to claim 16, wherein the event is a sports event.

18. The system according to claim 17, wherein said sports event includes a soccer game.

19. The system according to claim 16, wherein said multiple-layer models include a high level domain-specific knowledge model and a dynamic hierarchical event model.

20. The system according to claim 19, wherein said high level domain-specific knowledge model includes rules of a sports game.

21. The system according to claim 19, wherein said dynamic hierarchical event model includes a hierarchical decision tree.

22. The system according to claim 19, wherein said dynamic hierarchical event model includes an entity-relationship-diagram.

23. The system according to claim 16, wherein said at least one data source is selected from the group consisting of a camera, a microwave sensor, a sound recorder, and an input data stream.

24. The system according to claim 16, wherein said at least one data source includes a data stream sent through a network connection.

25. The system according to claim 24, wherein the data stream is a video stream with synchronized audio track.

26. The system according to claim 16, wherein the observation collection unit is simultaneously connected to more than one data source.

27. The system according to claim 16, wherein said temporal event detection unit includes an integration unit, a detection unit, and a fusion unit, and said integration unit combines a plurality of observation streams from a plurality of data sources, the detection unit detects a same event using a plurality of detection means to produce a plurality of detection results, and said fusion unit fuses the plurality of detection results to produce a single detection decision.

28. The system according to claim 16, further comprising:

an event storage, in which detected occurrences of temporal semantic events are stored;

an events statistics extractor to compute statistical information about the detected occurrences; and

an event statistics storage unit to store the statistical information.