Image matching apparatus, image matching method, computer program and computer-readable storage medium

• US 7,995,805 B2
• Filed: 08/29/2007
• Issued: 08/09/2011
• Est. Priority Date: 08/30/2006
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Suzuki, Takashi, Matsugu, Masakazu
• Primary Examiner(s): Mariam; Daniel G
• Application Number:
  US11/847,317
• Publication Number:
  US 20080219516A1
• Time in Patent Office: 1,441 Days
• Field of Search: 382/118, 382/209, 382/218, 382/224, 382/197, 382/274, 382/300
• US Class Current: 382/118
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation

Abstract

An image matching apparatus comprising, a detector adapted to detects, from an input image containing an object to be matched, a region where the object exists, an estimation unit adapted to estimates a parameter regarding a displaying characteristic based on an image of the region, a correction unit adapted to corrects the displaying characteristic of the input image based on the parameter, and a matching unit adapted to matches the corrected input image containing the object to be matched with an image containing a matching object.

Claims

1. An image matching apparatus comprising:

a computer-readable memory constructed to store computer-executable process steps; and

a processor constructed to execute the computer-executable process steps stored in the memory;

wherein the process steps stored in the memory cause the processor to perform;

a detector step of detecting, from an input image containing an object to be matched, a region where the object exists;

an estimation step of estimating a parameter regarding a luminance value based on an image of the region;

a correction step of correcting the luminance value of the input image based on the parameter; and

a matching step of matching the corrected input image containing the object to be matched with an image containing a matching object,

wherein, said correction step corrects the luminance value of the image by;

changing a luminance value smaller than a first threshold value to a minimum luminance value,

changing a luminance value larger than a second threshold value to a maximum luminance value, and

linearly interpolating a luminance value located between the first threshold value and the second threshold value.

2. The apparatus according to claim 1, wherein the process steps stored in the memory further cause the processor to perform:

a calculation step of calculating an area of a distribution of detection output related to the region; and

a determination step of determining whether the area falls within a predetermined range,

wherein said correction step corrects the luminance value of the input image when said determination step determines that the area does not fall within the predetermined range.

3. The apparatus according to claim 1, wherein,

the object to be matched is a human face, and wherein

said detection step comprises:

a first detection step of detecting a region where eyes and a mouth of the face are located; and

a second detection step of detecting, based on the region where the eyes and the mouth are located, a region where the face is located.

4. An image matching apparatus comprising:

a computer-readable memory constructed to store computer-executable process steps; and

a processor constructed to execute the computer-executable process steps stored in the memory;

wherein the process steps stored in the memory cause the processor to perform;

a detector step of detecting, from an input image containing an object to be matched, a region where the object exists;

an estimation step of estimating a parameter regarding a displaying characteristic based on an image of the region;

a correction step of correcting the displaying characteristic of the input image based on the parameter; and

a matching step of matching the corrected input image containing the object to be matched with an image containing a matching object,

wherein said matching step comprises;

an extraction unit step of extracting a specific region from the corrected input image;

a vector generation step of generating a feature vector based on the displaying characteristic of the specific region; and

a vector matching step of executing matching by a support vector machine using the feature vector and matching information registered in a registration unit.

5. The apparatus according to claim 4, wherein the specific region corresponds to eyes and a mouth when the object to be matched is a human.

6. An image matching method comprising:

using a computer to perform a detecting step of detecting, from an input image containing an object to be matched, a region where the object exists;

estimating a parameter regarding a luminance value based on an image of the region;

correcting the luminance value of the input image based on the parameter; and

matching the corrected input image containing the object to be matched with an image containing a matching object,

wherein the luminance value of the image is corrected by;

changing a luminance value smaller than a first threshold value to a minimum luminance value,

changing a luminance value larger than a second threshold value to a maximum luminance value, and

linearly interpolating a luminance value located between the first threshold value and the second threshold value.

7. The method according to claim 6, further comprising the steps of:

calculating an area of a distribution of detection output related to the region; and

determining whether the area falls within a predetermined range,

wherein the luminance value of the input image is corrected when it is determined in the determining step that the area does not fall within the predetermined range.

8. The method according to claim 6, wherein

the object to be matched is a human face, and wherein

said detecting the region comprises:

detecting a region where eyes and a mouth of the face are located; and

detecting, based on the region where the eyes and the mouth are located, a region where the face is located.

9. An image matching method comprising:

using a computer to perform a detecting step of detecting, from an input image containing an object to be matched, a region where the object exists;

estimating a parameter regarding a displaying characteristic based on an image of the region;

correcting the displaying characteristic of the input image based on the parameter; and

matching the corrected input image containing the object to be matched with an image containing a matching object, wherein said matching the corrected input image with the image containing the matching object comprises;

extracting a specific region from the corrected input image;

generating a feature vector based on a luminance value of the specific region; and

executing matching by a support vector machine using the feature vector and matching information.

10. The method according to claim 9, wherein the specific region corresponds to eyes and a mouth when the object to be matched is a human.

11. A non-transitory computer-readable storage medium retrievably storing a computer program for causing a computer to function as an image matching apparatus which comprises:

a detector constructed to detect, from an input image containing an object to be matched, a region where the object exists;

an estimation unit constructed to estimate a parameter regarding a luminance value based on an image of the region;

a correction unit constructed to correct the luminance value of the input image based on the parameter; and

a matching unit constructed to match the corrected input image containing the object to be matched with an image containing a matching object,

wherein, said correction unit corrects the luminance value of the image by;

changing a luminance value smaller than a first threshold value to a minimum luminance value,

changing a luminance value larger than a second threshold value to a maximum luminance value, and

linearly interpolating a luminance value located between the first threshold value and the second threshold value.

12. A non-transitory computer-readable storage medium retrievably storing a computer program for causing a computer to function as an image matching apparatus which comprises:

a detector constructed to detect, from an input image containing an object to be matched, a region where the object exists;

an estimation unit constructed to estimate a parameter regarding a displaying characteristic based on an image of the region;

a correction unit constructed to correct the displaying characteristic of the input image based on the parameter; and

a matching unit constructed to match the corrected input image containing the object to be matched with an image containing a matching object,

wherein said matching unit comprises;

an extraction unit constructed to extract a specific region from the corrected input image;

a vector generation unit constructed to generate a feature vector based on the displaying characteristic of the specific region; and

a vector matching unit constructed to execute matching by a support vector machine using the feature vector and matching information registered in a registration unit.