Robust object tracking system

• US 8,050,453 B2
• Filed: 06/15/2006
• Issued: 11/01/2011
• Est. Priority Date: 06/15/2006
• Status: Active Grant

• Current Assignee: Northwestern University, Omron Corporation
• Original Assignee: Northwestern University, Omron Corporation
• Inventors: Wu, Ying, Yang, Ming, Dai, Shengyang, Lao, Shihong
• Primary Examiner(s): Le; Brian Q
• Assistant Examiner(s): Bayat; Ali
• Application Number:
  US11/453,549
• Publication Number:
  US 20070291984A1
• Time in Patent Office: 1,965 Days
• Field of Search: 382/100, 382/103, 382/190, 382/236, 382/305, 348/169
• US Class Current: 382/103
• CPC Class Codes: G06T 7/20 Analysis of motion motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/43, H04N19/51

Abstract

A method for tracking objects includes identifying a target, identifying a plurality of auxiliary objects related to the target, and tracking the target using the plurality of auxiliary objects.

Claims

1. A method for tracking objects, comprising:

identifying a target;

identifying a plurality of auxiliary objects related to the target; and

tracking the target using the plurality of auxiliary objects,

wherein a data mining technique is used to identify the plurality of auxiliary objects related to the target, the data mining technique further comprising;

using a frequent pattern growth (FP growth) technique to select a plurality of candidate auxiliary objects with high co-occurrent frequency with the target; and

selecting the plurality of auxiliary objects from the plurality of candidate auxiliary objects,

wherein each of the plurality of auxiliary objects maintains a high motion correlation with the target.

2. The method of claim 1, further comprising using a mean shift technique to:

track the plurality of candidate auxiliary objects; and

determine motion correlations between the plurality of candidate auxiliary objects and the target.

3. The method of claim 2, wherein using the frequent pattern growth technique further comprises removing a candidate auxiliary object from the plurality of candidate auxiliary objects if the candidate auxiliary object cannot be tracked for a number of frames.

4. The method of claim 3, wherein the number of frames is four.

5. A tracking system configured to:

receive an image sequence;

identify a target in the image sequence;

identify a plurality of auxiliary objects related to the target; and

track the target using the plurality of auxiliary objects,

wherein a data mining technique is used to identify the plurality of auxiliary objects related to the target, the data mining technique further comprising;

using a frequent pattern growth (FP growth) technique to select a plurality of candidate auxiliary objects with high co-occurrent frequency with the target; and

selecting the plurality of auxiliary objects from the plurality of candidate auxiliary objects,

wherein each of the plurality of auxiliary objects maintains a high motion correlation with the target.

6. The tracking system of claim 5, further comprising

using a mean shift technique to:

track the plurality of candidate auxiliary objects; and

determine motion correlations between the plurality of candidate auxiliary objects and the target.

7. The tracking system of claim 6, wherein using the frequent pattern growth technique further comprises removing a candidate auxiliary object from the plurality of candidate auxiliary objects if the candidate auxiliary object cannot be tracked for a number of frames.

8. The tracking system of claim 6, wherein the number of frames is four.