Abandoned object recognition using pedestrian detection

• US 8,675,917 B2
• Filed: 10/31/2011
• Issued: 03/18/2014
• Est. Priority Date: 10/31/2011
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Brown, Lisa M., Feris, Rogerio S., Kjeldsen, Frederik C., Scherbaum, Kristina
• Primary Examiner(s): AZARIAN, SEYED H
• Application Number:
  US13/285,354
• Publication Number:
  US 20130108102A1
• Time in Patent Office: 869 Days
• Field of Search: 382/100, 382/103, 382106-107, 382115-118, 382/123, 382/155, 382/162, 382/168, 382/173, 382/181, 382189-195, 382/199, 382/203, 382/209, 382/224, 382/232, 382/254, 382/274, 382/276, 382/278, 382/291, 382/305, 382/312, 348/169, 735/143.2
• US Class Current: 382/103
• CPC Class Codes: G06F 18/24765 Rule-based classification
• CPC Class Codes: G06V 10/765 using rules for classification or partitioning the feature space
• CPC Class Codes: G06V 30/2504 Coarse or fine approaches, e.g. resolution of ambiguities or multiscale approaches
• CPC Class Codes: G06V 40/103 Static body considered as a whole, e.g. static pedestrian or occupant recognition

Abstract

Methods and apparatus are provided for improved abandoned object recognition using pedestrian detection. An abandoned object is detected in one or more images by determining if one or more detected objects in a foreground of the images comprises a potential abandoned object; applying a trained pedestrian detector to the potential abandoned object to determine if the potential abandoned object comprises at least a portion of a pedestrian; and classifying the potential abandoned object as an abandoned object based on whether the potential abandoned object is not at least a portion of a pedestrian. The trained pedestrian detector is trained using positive training samples comprised of at least portions of human bodies in one or more poses and/or negative training samples comprised of at least portions of abandoned objects.

Claims

1. A method for detecting an abandoned object in one or more images, comprising:

determining if one or more detected objects in a foreground of said images comprises a potential abandoned object;

applying a trained pedestrian detector to said potential abandoned object to determine if said potential abandoned object comprises at least a portion of a pedestrian, wherein said trained pedestrian detector is trained using positive training samples comprised of at least portions of human bodies in one or more poses; and

classifying said potential abandoned object as an abandoned object based on whether said potential abandoned object is not at least a portion of a pedestrian, wherein at least one of said determining, applying and classifying steps are performed by at least one hardware device.

2. The method of claim 1, wherein said trained pedestrian detector is trained using one or more of positive training samples and negative training samples.

3. The method of claim 2, wherein said negative training samples are comprised of at least portions of abandoned objects.

4. The method of claim 1, further comprising the step of determining whether said potential abandoned object has been put down by a person.

5. The method of claim 1, further comprising the step of determining whether said potential abandoned object has been left by a person.

6. The method of claim 1, further comprising the step of triggering an alarm if said potential abandoned object is classified as an abandoned object.

7. The method of claim 1, further comprising the step of processing said one or more images using a background subtraction (BGS) technique to detect moving objects in a stationary background.

8. The method of claim 1, further comprising the step of applying a healing technique to reduce foreground fragments.

9. The method of claim 1, wherein said determining step applies a mixture of Gaussians method to classify a background object as said abandoned object.

10. The method of claim 1, further comprising the step of determining if a size of said potential abandoned object meets predefined criteria for an abandoned object.

11. The method of claim 1, further comprising the step of determining if said potential abandoned object satisfies predefined abandoned time criteria.

12. The method of claim 1, further comprising the step of determining if an occlusion of said potential abandoned object in said images exceeds a predefined threshold, t.

13. An apparatus for detecting an abandoned object in one or more images, said apparatus comprising:

a memory; and

at least one hardware device, coupled to the memory, operative to;

determine if one or more detected objects in a foreground of said images comprises a potential abandoned object;

apply a trained pedestrian detector to said potential abandoned object to determine if said potential abandoned object comprises at least a portion of a pedestrian, wherein said trained pedestrian detector is trained using positive training samples comprised of at least portions of human bodies in one or more poses; and

classify said potential abandoned object as an abandoned object based on whether said potential abandoned object is not at least a portion of a pedestrian.

14. The apparatus of claim 13, wherein said trained pedestrian detector is further trained using negative training samples comprised of at least portions of abandoned objects.

15. The apparatus of claim 13, wherein said at least one hardware device is further configured to determine whether said potential abandoned object has been put down by a person.

16. The apparatus of claim 13, wherein said at least one hardware device is further configured to determine whether said potential abandoned object has been left by a person.

17. The apparatus of claim 13, wherein said at least one hardware device is further configured to trigger an alarm if said potential abandoned object is classified as an abandoned object.

18. The article of manufacture of claim 13, further comprising the step of triggering an alarm if said potential abandoned object is classified as an abandoned object.

19. An article of manufacture for detecting an abandoned object in one or more images, comprising a tangible machine readable recordable storage medium containing one or more programs which when executed implement the step of:

determining if one or more detected objects in a foreground of said images comprises a potential abandoned object;

applying a trained pedestrian detector to said potential abandoned object to determine if said potential abandoned object comprises at least a portion of a pedestrian, wherein said trained pedestrian detector is trained using positive training samples comprised of at least portions of human bodies in one or more poses; and

classifying said potential abandoned object as an abandoned object based on whether said potential abandoned object is not at least a portion of a pedestrian.

20. The article of manufacture of claim 19, wherein said trained pedestrian detector is further trained using negative training samples comprised of at least portions of abandoned objects.

21. The article of manufacture of claim 19, further comprising the step of determining whether said potential abandoned object has been put down by a person.

22. The article of manufacture of claim 19, further comprising the step of determining whether said potential abandoned object has been left by a person.