Optimizing the detection of objects in images

• US 9,405,981 B2
• Filed: 12/01/2015
• Issued: 08/02/2016
• Est. Priority Date: 10/20/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Li, Ying, Pankanti, Sharathchandra
• Primary Examiner(s): Bali, Vikkram
• Application Number:
  US14/955,342
• Publication Number:
  US 20160092741A1
• Time in Patent Office: 245 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: B60W 2554/00 Input parameters relating to objects
• CPC Class Codes: B60W 2556/50 of positioning data, e.g. GPS [Global Positioning System] data
• CPC Class Codes: G06T 2207/30261 Obstacle
• CPC Class Codes: G06T 7/70 Determining position or orientation of objects or cameras camera calibration G06T7/80
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
• CPC Class Codes: G06V 20/582 of traffic signs
• CPC Class Codes: H04N 23/56 provided with illuminating means
• CPC Class Codes: H04N 23/74 by influencing the scene brightness using illuminating means
• CPC Class Codes: H04N 7/185 from a mobile camera, e.g. for remote control

Abstract

A method, system, and computer program product, for detecting objects of interest in a digital image. At least positional data associated with a vehicle is received. Geographical information associated with the positional data is received. A probability of detecting an object of interest within a corresponding geographic area associated with the vehicle is determined based on the geographical data. The probability is compared to a given threshold. An object detection process is at least one of activated and maintained in an activated state in response to the probability being one of above and equal to the given threshold. The object detection process detects objects of interest within at least one image representing at least one frame of a video sequence of an external environment.

Claims

1. A method, with an information processing system, for detecting objects in a digital image, the method comprising:

receiving at least positional data associated with a vehicle;

obtaining geographical information associated with the positional data;

determining, based on the geographical information, a probability of detecting an object of interest within a corresponding geographic area associated with the vehicle;

comparing the probability to a given threshold;

in response to the probability being one of above and equal to the given threshold, performing the following;

receiving at least one image representing at least one frame of a video sequence of an external environment of the vehicle;

identifying a set of environmental conditions associated with the geographic area;

selecting an illumination model based on the set of environmental conditions;

retrieving at least one object template associated with the object of interest;

selecting an object appearance model based on the set of environmental conditions;

processing the at least one object template with respect to the illumination model and the object appearance model;

analyzing the at least one frame of a video sequence with respect to the at least one object template that has been processed; and

determining, based on the analyzing, if the at least one frame of a video sequence comprises the object of interest.

2. The method of claim 1, wherein the geographical information comprises a map.

3. The method of claim 1, wherein determining the probability of detecting the object of interest comprises:

analyzing the geographical information; and

determining, based on the analyzing, an environmental context associated with the geographic area, wherein the environmental context indicates at least a type of geographic area in which the vehicle is located, wherein the probability of detecting the object of interest is based on the environmental context.

4. The method of claim of claim 1, wherein determining the probability of detecting the object of interest comprises:

analyzing the geographical information with respect to the geographic area; and

determining, based on the analyzing, if the geographical information comprises at least one feature associated with the object of interest.

5. The method of claim 1, further comprising:

processing the at least one image with respect to the illumination model; and

analyzing the at least one image that has been processed with respect to the at least one object template that has been processed; and

wherein the determining comprises;

determining, based on the analyzing, if the at least one image that has been processed comprises the object of interest.

6. The method of claim 1, wherein the set of environmental conditions comprises at least one of:

a weather condition associated with the geographic area;

a time-of-day associated with the geographic area;

a current season associated with the geographic area; and

a lighting condition associated with the geographic area.

7. The method of claim 1, wherein the processing the at least one object template comprises at least one of:

normalizing a set of illumination characteristics of the at least one object template;

generating an illumination invariant object template;

generating an illumination specific object template from the at least one object template;

generating an illumination synthesized object template from the at least one object template; and

generating an appearance-adapted object template.

8. The method of claim 1, wherein the illumination model is one of an illumination invariant model, an illumination specific model, and a synthesized illumination model that is based on at least one other illumination model associated with a different set of environmental conditions.

9. An information processing system for detecting objects in a digital image, the information processing system comprising:

a memory;

a processor communicatively coupled to the memory; and

an object detection system communicatively coupled to the memory and the processor, the object detection system configured to perform a method comprising;

receiving at least positional data associated with a vehicle;

obtaining geographical information associated with the positional data;

determining, based on the geographical information, a probability of detecting an object of interest within a geographic area associated with the vehicle;

comparing the probability to a given threshold;

in response to the probability being one of above and equal to the given threshold, performing the following;

receiving at least one image representing at least one frame of a video sequence of an external environment of the vehicle;

identifying a set of environmental conditions associated with the geographic area;

selecting an illumination model based on the set of environmental conditions;

retrieving at least one object template associated with the object of interest;

selecting an object appearance model based on the set of environmental conditions;

processing the at least one object template with respect to the illumination model and the object appearance model;

analyzing the at least one frame of a video sequence with respect to the at least one object template that has been processed; and

determining, based on the analyzing, if the at least one frame of a video sequence comprises the object of interest.

10. The information processing system of claim 9, wherein the determining the probability of detecting the object of interest comprises:

analyzing the geographical information; and

determining, based on the analyzing, an environmental context associated with the geographic area, wherein the environmental context indicates at least a type of geographic area in which the vehicle is located, wherein the probability of detecting the object of interest is based on the environmental context.

11. The information processing system of claim 9, wherein the determining the probability of detecting the object of interest comprises:

analyzing the geographical information with respect to the geographic area; and

determining, based on the analyzing, if the geographical information comprises at least one feature associated with the object of interest.

12. The information processing system of claim 9, wherein the object detection system being configured to perform the method further comprising:

processing the at least one image with respect to the illumination model; and

analyzing the at least one image that has been processed with respect to the at least one object template that has been processed; and

determining, based on the analyzing, if the at least one image that has been processed comprises the object of interest.

13. The information processing system of claim 9, wherein the set of environmental conditions comprises at least one of:

a weather condition associated with the geographic area;

a time-of-day associated with the geographic area;

a current season associated with the geographic area; and

a lighting condition associated with the geographic area.

14. A computer program product for detecting objects in a digital image, the computer program product comprising:

a non-transitory storage medium readable by a processing circuit and storing instructions for execution by the processing circuit for performing a method comprising;

receiving at least positional data associated with a vehicle;

obtaining geographical information associated with the positional data;

determining, based on the geographical information, a probability of detecting an object of interest within a geographic area associated with the vehicle;

comparing the probability to a given threshold;

in response to the probability being one of above and equal to the given threshold, performing the following;

receiving at least one image representing at least one frame of a video sequence of an external environment of the vehicle;

identifying a set of environmental conditions associated with the geographic area;

selecting an illumination model based on the set of environmental conditions;

retrieving at least one object template associated with the object of interest;

selecting an object appearance model based on the set of environmental conditions;

processing the at least one object template with respect to the illumination model and the object appearance model;

analyzing the at least one frame of a video sequence with respect to the at least one object template that has been processed; and

determining, based on the analyzing, if the at least one frame of a video sequence comprises the object of interest.

15. The computer program product of claim 14, wherein determining the probability of detecting the object of interest comprises:

analyzing the geographical information; and

determining, based on the analyzing, an environmental context associated with the geographic area, wherein the environmental context indicates at least a type of geographic area in which the vehicle is located, wherein the probability of detecting the object of interest is based on the environmental context.

16. The computer program product of claim 14, wherein the determining the probability of detecting the object of interest comprises:

analyzing the geographical information with respect to the corresponding geographic area; and

determining, based on the analyzing, if the geographical information comprises at least one feature associated with the object of interest.

17. The computer program product of claim 14, wherein the instructions for execution by the processing circuit for performing the method further comprising:

processing the at least one image with respect to the illumination model; and

analyzing the at least one image that has been processed with respect to the at least one object template that has been processed; and

determining, based on the analyzing, if the at least one image that has been processed comprises the object of interest.

18. The computer program product of claim 14, wherein the set of environmental conditions comprises at least one of:

a weather condition associated with the geographic area;

a time-of-day associated with the geographic area;

a current season associated with the geographic area; and

a lighting condition associated with the geographic area.

19. The computer program product of claim 14, wherein the processing the at least one object template comprises at least one of:

normalizing a set of illumination characteristics of the at least one object template;

generating an illumination invariant object template;

generating an illumination specific object template from the at least one object template;

generating an illumination synthesized object template from the at least one object template; and

generating an appearance-adapted object template.

20. The method of claim 14, wherein the illumination model is one of an illumination invariant model, an illumination specific model, and a synthesized illumination model that is based on at least one other illumination model associated with a different set of environmental conditions.