METHOD AND SYSTEM FOR OBJECT DETECTION IN IMAGES UTILIZING ADAPTIVE SCANNING

• US 20090245570A1
• Filed: 03/28/2008
• Published: 10/01/2009
• Est. Priority Date: 03/28/2008
• Status: Active Grant

• Current Assignee: Honeywell International Inc.
• Original Assignee: Honeywell International Inc.
• Inventors: Juza, Michal, Gold, Tomas
• Granted Patent: US 8,538,171 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06V 10/25 Determination of region of interest [ROI] or a volume of interest [VOI]

Abstract

An object detection method and system for detecting an object in an image utilizing an adaptive image scanning strategy is disclosed herein. An initial rough shift can be determined based on the size of a scanning window and the image can be scanned continuously for several detections of similar sizes using the rough shift. The scanning window can be classified with respect to a cascade of homogenous classification functions covering one or more features of the object. The size and scanning direction of the scanning window can be adaptively changed depending on the probability of the object occurrence in accordance with scan acceleration. The object can be detected by an object detector and can be localized with higher precision and accuracy.

Claims

1. A computer implemented method for detecting objects in an image, comprising:

determining an initial rough shift of an input image based on a size of a scanning window in order to scan said input image for a plurality of potential detections of similar sizes utilizing said initial rough shift, wherein said scanning window is classifiable with respect to a cascade of homogenous classification functions;

adaptively changing a size and a scanning direction of said scanning window with respect to said cascade of homogenous classification functions depending on a probability of an object occurrence in a particular area of said scanning window; and

detecting said object associated with said scanning window based on a response from an object detector in order to thereafter localize said object having sufficient features with a higher precision and accuracy thereto.

2. The computer implemented method of claim 1 further comprising:

determining a set of sizes associated with said scanning window with respect to said input image size.

3. The computer implemented method of claim 1 further comprising:

estimating at least one gradient of a confidence function by shifting said scanning window in all directions utilizing a small step and a small variation in size of said scanning window wherein said confidence function comprises an output of said object detector; and

outputting said object detector result indicating the presence of said object in said scanning window on said input image.

4. The computer implemented method of claim 1 wherein said cascade of homogenous classification functions and said input image are stored in a database.

5. The computer implemented method of claim 1 wherein said object comprises a particular feature of said image.

6. The computer implemented method of claim 5 wherein said particular feature of said image comprises a head of a person exhibiting a facial region.

7. The computer implemented method of claim 1 wherein an output of said object detector comprises a measure of correct object alignment.

8. A system, comprising:

a processor;

a data bus coupled to said processor; and

a computer-usable medium embodying computer code, said computer-usable medium being coupled to said data bus, said computer program code comprising instructions executable by said processor and configured for;

determining an initial rough shift of an input image based on a size of a scanning window in order to scan said input image for a plurality of potential detections of similar sizes utilizing said initial rough shift wherein said scanning window is classifiable with respect to a cascade of homogenous classification functions;

adaptively changing a size and a scanning direction of said scanning window with respect to said cascade of homogenous classification functions depending on a probability of an object occurrence in a particular area of said scanning window; and

detecting said object associated with said scanning window based on a response from an object detector in order to thereafter localize said object having sufficient features with a higher precision and accuracy thereto.

9. The system of claim 8 wherein said instructions are further configured for:

determining a set of sizes associated with said scanning window with respect to said input image size.

10. The system of claim 8 wherein said instructions are further configured for:

estimating at least one gradient of a confidence function by shifting said scanning window in all directions utilizing a small step and a small variation in size of said scanning window wherein said confidence function comprises an output of said object detector; and

outputting said object detector result indicating the presence of said object in said scanning window on said input image.

11. The system of claim 8 wherein said cascade of homogenous classification functions and said input image are stored in a database.

12. The system of claim 9 wherein said object comprises a particular feature of said image.

13. The system of claim 8 wherein said instructions are further configured for:

determining a set of sizes associated with said scanning window with respect to said input image size, wherein said object comprises an object for which said object detector is trained to identify.

14. The system of claim 13 wherein said cascade of homogenous classification functions and said input image are stored in a database.

15. A computer-usable medium embodying computer program code, said computer program code comprising computer executable instructions configured for:

determining an initial rough shift of an input image based on a size of a scanning window in order to scan said input image for a plurality of potential detections of similar sizes utilizing said initial rough shift wherein said scanning window is classifiable with respect to a cascade of homogenous classification functions;

adaptively changing a size and a scanning direction of said scanning window with respect to said cascade of homogenous classification functions depending on a probability of an object occurrence in a particular area of said scanning window; and

detecting said object associated with said scanning window based on a response from an object detector in order to thereafter localize said object having sufficient features with a higher precision and accuracy thereto.

16. The computer-usable medium of claim 15, wherein said embodied computer program code further comprises computer executable instructions configured for:

determining a set of sizes associated with said scanning window with respect to said input image size.

17. The computer-usable medium of claim 15, wherein said embodied computer program code further comprises computer executable instructions configured for:

estimating at least one gradient of a confidence function by shifting said scanning window in all directions utilizing a small step and a small variation in size of said scanning window wherein said confidence function comprises an output of said object detector; and

outputting said object detector result indicating the presence of said object in said scanning window on said input image.

18. The computer-usable medium of claim 15, wherein said cascade of homogenous classification functions and said input image are stored in a database.

19. The computer-usable medium of claim 15, wherein said object comprises a particular feature of said image.

20. The computer-usable medium of claim 15, wherein said embodied computer program code further comprises computer executable instructions configured for:

determining a set of sizes associated with said scanning window with respect to said input image size, wherein said object comprises a head of a person exhibiting a facial region.