Feature Change Image Creation Method, Feature Change Image Creation Device, and Feature Change Image Creation Program

• US 20080240489A1
• Filed: 01/06/2005
• Published: 10/02/2008
• Est. Priority Date: 01/13/2004
• Status: Abandoned Application

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Marugame, Atsushi
• Application Number:
  US10/597,148
• Publication Number:
  US 20080240489A1
• Time in Patent Office: Days
• US Class Current: 382/100
• CPC Class Codes: G06T 1/00 General purpose image data processing

Abstract

A technique for applying a feature of a side attribute to an original image while keeping a principal feature. An image accumulating unit stores images classified into categories per the age or sex. A merging unit selects a category in accordance with an input category. A matching unit matches the input image and each of images included in the selected category with each other, and selects the image having highest degree of similarity. The merging unit merges the input image and the selected image by carrying out the weighted average and so on, and generates and outputs a merged image.

Claims

1. A feature changed image generating method for generating a new image from an input image, comprising:

providing a database in which a plurality of data, which are relating to a plurality of images respectively, are classified into a plurality of categories;

determining an image which is most similar to said input image as a selected image based on a data belonging to a specified category specified from said plurality of categories; and

merging said selected image and said input image.

2. The feature changed image generating method according to claim 1,

wherein a database in which said plurality of images are classified into said plurality of categories is provided in said providing, and

an image which is most similar to said input image among images belonging to said specified category is selected as said selected image in said determining.

3. The feature changed image generating method according to claim 1,

wherein a database in which constituent components of said plurality of images are classified into said plurality of categories is provided in said providing, and

said determining includes:

determining a determined combination of said constituent components by which an image which is most similar to said input image is obtained by using said constituent components belonging to said specified category; and

generating an image which is most similar to said input image as said selected image based on said determined combination.

4. The feature changed image generating method according to claim 1,

wherein a database in which said plurality of images are classified into said plurality of categories is provided, and each of said plurality of categories includes a plurality of images which are gradual variations of an identical object on an attribute, and

said determining includes:

selecting an image which is most similar to said input image among images belonging to a category included in said plurality of categories and corresponding to an attribute of said input image as a similar image; and

determining an image relating to a same object with said similar image as said selected image from images belonging to said specified category.

5. The feature changed image generating method according to claim 1,

wherein a database in which constituent components of said plurality of images are classified into said plurality of categories is provided, and each of said plurality of categories includes constituent components of a plurality of images which are gradual variations of an identical object on an attribute, and

said determining includes:

selecting a selected combination of said constituent components by which an image which is most similar to said input image is obtained, by using said constituent components belonging to a category included in said plurality of categories and corresponding to an attribute of said input image;

converting component coefficients corresponding to said selected combination into converted coefficients which are component coefficients corresponding to said specified category; and

generating said selected image by using said converted coefficients and said constituent components belonging to said specified category.

6. The feature changed image generating method according to claim 1,

wherein each of said plurality of images is a face image of a person, and

said plurality of categories are categorized based on an age.

7. The feature changed image generating method according to claim 6,

wherein a category included in said plurality of categories and corresponding to an age higher than said specified age is selected as said specified category when an age of a person on said input image is lower than an age specified by a user.

8. The feature changed image generating method according to claim 6,

wherein a category included in said plurality of categories and corresponding to an age lower than said specified age is selected as said specified category when an age of a person in said input image is higher than an age specified by a user.

9. A feature change applying method for gradually applying a feature change to an input image, comprising:

providing a database in which constituent components of a plurality of images are classified into a plurality of categories, wherein each of said plurality of categories includes constituent components of a plurality of images which are gradual variations of an identical object on an attribute;

selecting a selected combination of said constituent components by which an image which is most similar to said input image is obtained, by using said constituent components belonging to a category included in said plurality of categories and corresponding to an attribute of said input image; and

converting component coefficients corresponding to said selected combination into converted coefficients which are component coefficients corresponding to said specified category.

10. The feature change applying method according to claim 9,

wherein each of said plurality of images is a face image of a person, and

said plurality of categories are categorized by based on an age.

11. A feature changed image generating apparatus for generating a new image from an input image, comprising:

a storing unit configured to store a plurality of data which are relating to a plurality of images respectively and classified into a plurality of categories;

an image determining unit configured to determine an image which is most similar to said input image as a selected image based on a data belonging to a specified category specified from said plurality of categories; and

a merging unit configured to merge said selected image and said input image.

12. The feature changed image generating apparatus according to claim 11,

wherein said plurality of images are classified into said plurality of categories in said storing unit, and

said image determining unit determines an image which is most similar to said input image among images belonging to said specified category as said selected image.

13. The feature changed image generating apparatus according to claim 11,

wherein a constituent components of said plurality of images are classified into said plurality of categories in said storing unit, and

said image determining unit determines a determined combination of said constituent components by which an image which is most similar to said input image as said selected image based on said determined combination.

14. The feature changed image generating apparatus according to claim 11,

wherein said storing unit stores said plurality of images classified into said plurality of categories, and each of said plurality of categories includes a plurality of images which are gradual variations of an identical object on an attribute, and

said image determining unit selects an image which is most similar to said input image among images belonging to a category included in said plurality of categories and corresponding to an attribute of said input image as a similar image, and determines an image relating to a same object with said similar image as said selected image from images belonging to said specified category.

15. The feature changed image generating apparatus according to claim 11,

wherein constituent components of said plurality of images are classified into said plurality of categories in said storing unit, and each of said plurality of categories includes constituent components of a plurality of images which are gradual variations of an identical object on an attribute, and

said image determining unit selects a selected combination of said constituent components by which an image which is most similar to said input image is obtained, by using said constituent components belonging to a category included in said plurality of categories and corresponding to an attribute of said input image, converts component coefficients corresponding to said selected combination into converted coefficients which are component coefficients corresponding to said specified category, and generates said selected image by using said converted coefficients and said constituent components belonging to said specified category.

16. The feature changed image generating apparatus according to claim 11,

wherein each of said plurality of images is a face image of a person, and

said plurality of categories are categorized based on an age.

17. The feature changed image generating apparatus according to claim 16, further comprising a selecting unit,

wherein said selecting unit selects a category included in said plurality of categories and corresponding to an age higher than said specified age as said specified category when an age of a person on said input image is lower than an age specified by a user.

18. The feature changed image generating apparatus according to claim 16, further comprising a selecting unit,

wherein said selecting unit selects a category included in said plurality of categories and corresponding to an age lower than said specified age as said specified category when an age of a person on said input image is higher than an age specified by a user.

19. A feature change applying apparatus for gradually applying a feature change to an input image, comprising:

a storing unit in which constituent components of a plurality of images are classified into a plurality of categories; and

a component coefficient converting unit,

wherein each of said plurality of categories includes constituent components of a plurality of images which are gradual variations of an identical object on an attribute, and

said component coefficient converting unit selects a selected combination of said constituent components by which an image which is most similar to said input image is obtained, by using said constituent components belonging to a category included in said plurality of categories and corresponding to an attribute of said input image, and converts component coefficients corresponding to said selected combination into converted coefficients which are component coefficients corresponding into said specified category.

20. The feature change applying apparatus according to claim 19,

wherein each of said plurality of images is a face image of a person, and

said plurality of categories are categorized based on an age.

21. A feature changed image generating program for generating a new image from an input image executed by a computer, comprising a storing device storing a plurality of data which are relating to a plurality of images respectively and classified into a plurality of categories, and

the feature changed image generating program causes the computer to execute:

determining an image which is most similar to said input image as a selected image based on a data belonging to a specified category specified from said plurality of categories; and

merging said selected image and said input image.

22. The feature changed image generating program according to claim 21,

wherein said plurality of images are classified into said plurality of categories in said storing device, and

the feature changed image generating program causes the computer to execute determining an image which is most similar to said input image among images belonging to said specified category as said selected image.

23. The feature changed image generating program according to claim 21,

wherein constituent components of said plurality of images classified into said plurality of categories are stored in said storing device, and

the feature changed image generating program causes the computer to execute:

determining a determined combination of said constituent components by which an image which is most similar to said input image is obtained by using said constituent components belonging to said specified category; and

generating an image which is most similar to said input image as said selected image based on said determined combination.

24. The feature changed image generating program according to claim 21,

wherein said storing device stores said plurality of images classified into said plurality of categories, and each of said plurality of categories includes a plurality of images which are gradual variations of an identical object on an attribute, and

the feature changed image generating program causes the computer to execute:

selecting an image which is most similar to said input image among images belonging to a category included in said plurality of categories and corresponding to an attribute of said input image as a similar image; and

determining an image relating to a same object with said similar image as said selected image from images belonging to said specified category.

25. The feature changed image generating program according to claim 21,

wherein said storing device stores constituent components of said plurality of images classified into said plurality of categories, and each of said plurality of categories includes constituent components of a plurality of images which are gradual variations of an identical object on an attribute, and

said feature changed image generating program causes the computer to execute:

selecting a selected combination of said constituent components by which an image which is most similar to said input image is obtained, by using said constituent components belonging to a category included in said plurality of categories and corresponding to an attribute of said input image;

converting component coefficients corresponding to said selected combination into converted coefficients which are component coefficients corresponding to said specified category; and

generating said selected image by using said converted coefficients and said constituent components belonging to said specified category.

26. A feature change applying program for gradually applying a feature change to an input image executed by a computer,

wherein the computer has a storing device in which constituent components of a plurality of images are classified into a plurality of categories, and each of said plurality of categories includes constituent components of a plurality of images which are gradual variations of an identical object on an attribute, and

the feature change applying program causes the computer to execute:

selecting a selected combination of said constituent components by which an image which is most similar to said input image is obtained, by using said constituent components belonging to a category included in said plurality of categories and corresponding to an attribute of said input image; and

converting component coefficients corresponding to said selected combination into converted coefficients which are component coefficients corresponding to said specified category.