Template matching method, template matching apparatus, and recording medium that records program for it

• US 7,813,558 B2
• Filed: 01/11/2006
• Issued: 10/12/2010
• Est. Priority Date: 01/11/2005
• Status: Expired due to Fees

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Hatabu, Atsushi
• Primary Examiner(s): Ahmed; Samir A.
• Assistant Examiner(s): HU, FRED H.
• Application Number:
  US10/594,897
• Publication Number:
  US 20080273800A1
• Time in Patent Office: 1,735 Days
• Field of Search: None
• US Class Current: 382/209
• CPC Class Codes: G06T 7/74 involving reference images or patches
• CPC Class Codes: G06V 10/7515 Shifting the patterns to accommodate for positional errors
• CPC Class Codes: G06V 30/2504 Coarse or fine approaches, e.g. resolution of ambiguities or multiscale approaches

Abstract

There is provided a template matching method having a small amount of calculation and high detection accuracy. The amount of calculation is reduced while minimizing the degradation of the detection accuracy by calculating the similarity between a template image region and a reference image region at rough position intervals and estimating by interpolation of the similarity at a position where it is not calculated from the similarity calculated at a peripheral position. At the time, even if the gradient of a similarity greatly changes at the minimal point, estimation can be realized with high accuracy by executing estimation making use of the directional gradients of the similarity obtained in plural different directions independently. Additionally, occurrence of wrong detection is avoided by the restriction to disable the difference between the estimated similarity and adjacent similarity to exceed a threshold, which is defined based on the similarity between the template image region and the image region obtained by moving the template image region in the same direction as or in the opposite direction to the direction of estimation.

Claims

1. A template matching method of detecting a position of an image region similar to a template image region from on a reference image, said method comprising the steps of:

calculating a similarity of the image region to the template image region at rough position intervals;

estimating a similarity at a position, where the similarity is not calculated, making use of directional gradients of similarity obtained independently in a plurality of directions from the calculated values of similarity;

determining, based on the calculated values of similarity obtained in the calculating step and the estimated value of similarity obtained in the estimating step, a position having a small calculated or estimated value of the similarity as a position of a similar image region; and

restricting a value range of the estimated similarity value such that a difference between the estimated similarity value and the similarity of a periphery used for the estimation or the gradient of the similarity does not exceed a threshold value,

wherein the calculating step, estimating step, determining step and restricting step are performed by a processor.

2. A template matching method according to claim 1, further comprising detecting the position of the similar image region at a search step before the final step thereof.

3. A template matching method according to claim 1, further comprising a step of determining the threshold value based on a magnitude of the similarity calculated from the template image region and an image region, which is obtained by moving the template image region in a same direction or in a opposite direction to an estimation direction vector obtained by subtracting a position of a nearby image region from a position where the similarity is estimated.

4. A template matching method according to claim 1, further comprising a step of estimating, when five reference image regions A, B, C, D, E are sequentially located on a straight line, a degree of similarity at the image region C from an interpolation value extrapolated assuming gradient continuity from a similarity in the image regions A, B and from an interpolation value extrapolated assuming gradient continuity from a similarity in the image regions D, E in the estimation of the degree of similarity.

5. A template matching apparatus for detecting the position of an image region similar to a template image region from on a reference image, comprising:

means for calculating a similarity of the image region to the template image region at rough position intervals;

means for estimating a similarity at a position, where the similarity is not calculated, making use of gradients of the similarity obtained independently in a plurality of directions from the calculated values of similarity;

means for determining, based on the calculated values of similarity obtained by the means for calculating and the estimated value of similarity obtained by the means for estimating, a position having a small calculated or estimated value of the similarity as a position of a similar image region; and

means for restricting a value range of the estimated similarity value such that a difference between the estimated similarity value and the similarity of a periphery used for the estimation or the gradient of the similarity does not exceed a threshold value.

6. A template matching apparatus for narrowing down the position of a similar image region stepwise based on a multistep search method, further comprising detecting the position of the similar image region by the template matching apparatus according to claim 5 at a search step before the final step thereof.

7. A template matching apparatus according to claim 5, further comprising means for determining the threshold value based on the magnitude of the similarity calculated from the template image region and an image region, which is obtained by moving the template image region in a same direction as or in a opposite direction to an estimation direction vector obtained by subtracting the position of the nearby image region from the position where the degree of similarity is estimated.

8. A template matching apparatus according to claim 5, further comprising means for estimating, when five reference image regions A, B, C, D, E are sequentially located on a straight line, a degree of similarity at the image region C from an interpolation value extrapolated assuming gradient continuity from a similarity in the image regions A, B and from an interpolation value extrapolated assuming gradient continuity from a similarity in the image regions D, E in the estimation of the degree of similarity.

9. A non-transitory computer readable recording medium that records a program for causing a computer to execute a template matching method of detecting the position of an image region similar to a template image region from on a reference image, wherein:

the method calculates a similarity of the image region to the template image region at rough position intervals;

the method estimates a similarity at a position, where the similarity is not calculated, making use of gradients of similarity obtained independently in a plurality of directions from the calculated values of similarity;

the method determines, based on the calculated values of similarity obtained in the calculating step and the estimated value of similarity obtained in the estimating step, a position having a small calculated or estimated value of the similarity as a position of a similar image region; and

the method restricts a value range of the estimated similarity value such that a difference between the estimated similarity value and the similarity of a periphery used for the estimation or the gradient of the similarity does not exceed a threshold value.

10. A non-transitory computer readable recording medium that records a program for causing a computer to execute a template matching method of narrowing down the position of a similar image region stepwise based on a multistep search method, characterized in that the template matching method detects the position of the similar image region by being executed by a computer based on a program recorded on the recording medium according to claim 9 at a search step before the final step thereof.

11. A non-transitory recording medium according to claim 9, wherein the method determines the threshold value based on the magnitude of the similarity calculated from the template image region and an image region, which is obtained by moving the template image region in the same direction as or in the opposite direction to an estimation direction vector obtained by subtracting the position of the nearby image region from the position where the degree of similarity is estimated.

12. A non-transitory recording medium according to claim 9, wherein the method estimates, when five reference image regions A, B, C, D, E are sequentially located on a straight line, the degree of similarity at the position C from an interpolation value extrapolated assuming gradient continuity from the similarity in the image regions A, B and from an interpolation value extrapolated assuming gradient continuity from the similarity in the image regions D, E in the estimation of the degree of similarity.