Computer implemented system and method for high performance visual tracking

1. A computer system for identifying an activity, event, or chain of events in a video system, the computer system comprising:

a computer containing a plurality of software modules,

a video stream input, and

wherein the plurality of software modules include

a video capturing module;

an object tracking module;

a track identification module;

a periodic motion tracking module;

a periodic motion identifying module; and

a reasoning module;

wherein

the video capturing module receives images from a video stream and outputs image data directly to both the object tracking module and the periodic motion tracking module;

the object tracking module receives the image data from the video capturing module and produces a tracked path of an object in the video stream based on the image data; and

the track identification module receives the tracked path from the object tracking module and performs a comparison of the tracked path to a model;

the periodic motion tracking module receives the image data from the video capturing module and creates and outputs a data structure representing motion;

the periodic motion identifying module identifies periodic motion and non-periodic motion within the image data based on the output from the periodic motion tracking module; and

the reasoning module receives both the comparison from the track identification module and data regarding the identification of periodic motion and non-periodic motion from the periodic motion identifying module, and based thereon the reasoning module identifies motion characteristics of the activity, event, or chain of events.

2. The computer system according to claim 1 , herein the video stream input includes a camera configured to detect light in a visible spectrum and/or an infrared spectrum.

3. The computer system according to claim 1 , wherein the video stream input is configured to receive the video stream from a hardware memory having prerecorded video data stored thereon.

4. The computer system according to claim 1 , wherein the object tracking module

identifies the object in the video stream based on said received image data;

selects a tracking algorithm; and

tracks the object using the tracking algorithm and starts a track.

5. The computer system according to claim 1 , wherein the periodic motion tracking module extracts information from a stabilized foreground image and compares said extracted information to information extracted from prior stabilized foreground images to create the data structure representing motion.

6. The computer system according to claim 1 , wherein the periodic motion identifying module identifies periodic motion and non-periodic motion includes analyzing a similarity array and identifying periods of increasing and decreasing similarity between directional histograms.

7. The computer system according to claim 6 , wherein the periodic motion identifying module identifies periodic motion and non-periodic motion includes fitting one or more Gaussian distribution functions to data in the similarity array.

8. The computer system according to claim 1 , wherein the periodic motion identifying module identifies periodic motion and non-periodic motion includes mapping periodicity values to a periodic motion string using thresholds.

9. The computer system according to claim 1 , wherein the object tracking module is configured to detect an occlusion of the object and compensate for the detected occlusion.

10. The computer system according to claim 1 , wherein the reasoning module identifies motion characteristics of the activity, event, or chain of events based on a one or a combination of a determined probability image, a color histogram, a motion image, and an occlusion detection.

11. A method for identifying an activity, event, or chain of events in a video system, the method being implemented by a computer system containing a plurality of software modules and a video stream input, the method comprising:

receiving by a video capturing module images from a video stream and outputting by the video capturing module image data directly to both an object tracking module and a periodic motion tracking module;

receiving by the object tracking module the image data from the video capturing module and producing by the object tracking module a tracked path of an object in the video stream based on the image data;

receiving by a track identification module the tracked path from the object tracking module and performing by the track identification module a comparison of the tracked path to a model;

receiving by the periodic motion tracking module the image data from the video capturing module and creating and outputting by the periodic motion tracking module a data structure representing motion;

identifying by a periodic motion identifying module periodic motion and non-periodic motion within the image data based on the output from the periodic motion tracking module; and

receiving by a reasoning module both the comparison from the track identification module and data regarding the identification of periodic motion and non-periodic motion from the periodic motion identifying module, and based thereon identifying by the reasoning module motion characteristics of the activity, event, or chain of events.

12. The method according to claim 11 , wherein the video stream is received by a video stream input that includes a camera configured to detect light in a visible spectrum and/ or an infrared spectrum.

13. The method according to claim 11 , wherein the video stream is received from a hardware memory having prerecorded video data stored thereon.

14. The method according to claim 11 , wherein producing the tracked path by the object tracking module includes

identifying the object in the video stream based on said received image data;

selecting a tracking algorithm; and

tracking the object using the tracking algorithm and starts a track.

15. The method according to claim 11 , wherein the periodic motion tracking module extracts information from a stabilized foreground image and compares said extracted information to information extracted from prior stabilized foreground images to create the data structure representing motion.

16. The method according to claim 11 , wherein the periodic motion identifying module identifies periodic motion and non-periodic motion includes analyzing a similarity array and identifying periods of increasing and decreasing similarity between directional histograms.

17. The method according to claim 16 , wherein the periodic motion identifying module identifies periodic motion and non-periodic motion includes fitting one or more Gaussian distribution functions to data in the similarity array.

18. The method according to claim 11 , wherein the periodic motion identifying module identifies periodic motion and non-periodic motion includes mapping periodicity values to a periodic motion string using thresholds.

19. The method according to claim 11 , wherein the object tracking detects an occlusion of the object and compensates for the detected occlusion.

20. One or more non-transitory computer-readable media having stored thereon executable instructions that when executed by one or more processors of a computer system configure the computer system to perform at least the following:

receive by a video capturing module of the computer system images from a video stream and output by the video capturing module image data directly to both an object tracking module of the computer system and a periodic motion tracking module of the computer system;

receive by the object tracking module of the computer system the image data from the video capturing module and produce by the object tracking module a tracked path of an object in the video stream based on the image data;

receive by a track identification module of the computer system the tracked path from the object tracking module and perform by the track identification module a comparison of the tracked path to a model;

receive by the periodic motion tracking module the image data from the video capturing module, and create and output by the periodic motion tracking module a data structure representing motion;

identify by a periodic motion identifying module of the computer system periodic motion and non-periodic motion within the image data based on the output from the periodic motion tracking module; and

receive by a reasoning module of the computer system both the comparison from the track identification module and data regarding the identification of periodic motion and non-periodic motion from the periodic motion identifying module, and based thereon identify by the reasoning module motion characteristics of an activity, event, or chain of events.