Systems, methods, and devices for image matching and object recognition in images using feature point optimization
An image matching technique locates feature points in a template image such as a logo and then does the same in a test image. Feature points of a template image are determined under various transformations and used to determine a set of composite feature points for each template image. The composite feature points are used to determine if the template image is present in a test image.
1. A computer-implemented, implemented by hardware in combination with software, the method comprising:
(A) producing multiple template images from a particular template image, said multiple template images being produced by multiple transformations of said particular template image;
(B) for each of said multiple template images and for said particular template image determining a corresponding set of feature points;
(C) determining a composite set of feature points from feature points associated with each of said multiple template images, said composite set of feature points comprising invariant feature points selected from feature points associated with each of said multiple template images, wherein an invariant feature point is a feature point that is present in at least a predetermined number of template images; and
(D) associating said composite set of feature points with said particular template image,
wherein said determining in (C) comprises:
(C)(1) determining that feature points from distinct images of said multiple template images are similar using distance measures in Euclidean space and feature descriptor space.
2. The method of claim 1 wherein said determining is (C) also determines said composite set of feature points from feature points associated with said particular template image.
3. The method of claim 1 wherein said composite set of feature points comprises invariant feature points selected from: (i) feature points associated with each of said multiple template images, and (ii) feature points associated with said particular template image.
4. The method of claim 1 wherein said each of said multiple transformations comprises one or more transformations selected from: scaling, blurring, stretching, rotating, skewing, translating, distorting, contrast adjusting, color adjusting, morphological filtering, and projective transformations.
5. The method of claim 1 further comprising:
(E) using said composite set of feature points associated with said particular template image to determine whether at least a portion of said particular template image is present in a test image.
6. The method of claim 5 further comprising:
(F) based on said determining in (E), when it is determined using said composite set of feature points that said at least a portion of said particular template image is present in a test image, verifying that said at least a portion of said particular template image is present in a test image using said feature points associated with said particular template image.
7. An article of manufacture, comprising non-transitory computer-readable media having computer readable instructions stored thereon, the computer readable instructions including instructions for implementing a computer-implemented method, said method operable on one or more devices comprising hardware including memory and at least one processor and running one or more services on said hardware, said method comprising:
(A) producing multiple template images from a particular template image, said multiple template images being produced by multiple transformations of said particular template image;
(B) for each of said multiple template images and for said particular template image determining a corresponding set of feature points;
(C) determining a composite set of feature points from feature points associated with each of said multiple template images, said composite set of feature points comprising invariant feature points selected from feature points associated with each of said multiple template images, wherein an invariant feature point is a feature point that is present in at least a predetermined number of template images; and
(D) associating said composite set of feature points with said particular template image,
wherein said determining in (C) comprises:
(C)(1) determining that feature points from distinct images of said multiple template images are similar using distance measures in Euclidean space and feature descriptor space.
8. The article of manufacture of claim 7 wherein said determining is (C) also determines said composite set of feature points from feature points associated with said particular template image.
9. The article of manufacture of claim 7 wherein said composite set of feature points comprises invariant feature points selected from: (i) feature points associated with each of said multiple template images, and (ii) feature points associated with said particular template image.
10. The article of manufacture of claim 7 wherein said each of said multiple transformations comprises one or more transformations selected from: scaling, blurring, stretching, rotating, skewing, translating, distorting, contrast adjusting, color adjusting, morphological filtering, and projective transformations.
11. The article of manufacture of claim 7 wherein the method further comprises:
(E) using said composite set of feature points associated with said particular template image to determine whether at least a portion of said particular template image is present in a test image.
12. The article of manufacture of claim 7 wherein the method further comprises:
(F) based on said determining in (E), when it is determined using said composite set of feature points that said at least a portion of said particular template image is present in a test image, verifying that said at least a portion of said particular template image is present in a test image using said feature points associated with said particular template image.