Object extraction method, and image sensing apparatus using the method

• US 6,636,635 B2
• Filed: 10/31/1996
• Issued: 10/21/2003
• Est. Priority Date: 11/01/1995
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Matsugu, Masakazu
• Primary Examiner(s): Johnson, Timothy M.
• Application Number:
  US08/742,206
• Publication Number:
  US 20020044691A1
• Time in Patent Office: 2,546 Days
• Field of Search: 382/164, 382/165, 382/170, 382/171, 382/205, 382/209, 382/217, 382/218, 382/219, 382/220, 382/274, 382/282, 382/283, 382/294, 382/312, 382/215, 382/216, 382/199, 348/129, 348/161, 348/225, 348/345, 348/352, 348/364, 348/365, 348/366, 348/370, 358/464, 358/453, 345/433, 345/434, 345/523
• US Class Current: 382/218
• CPC Class Codes: G06V 10/10 Image acquisition document image scanning and transmission H04N1/00; control of digital cameras H04N23/60

Abstract

This invention provides an object extraction method for performing processing for extracting and cutting out a specific object from a sensed image at high speed, and an image sensing apparatus using the method. In this invention, in a method of extracting an object by comparing a sensed image and a standard image, a focusing signal, focal length data, visual axis direction data, and illumination conditions are detected, and the initial size, initial position, or initial color of the standard image is changed on the basis of the detection results, and extraction is started under optical conditions.

Claims

1. An object extraction method for extracting a predetermined object from a sensed image, comprising the steps of:

automatically changing a shape and/or a size of reference image associated with object extraction on the basis of at least an image sensing parameter of an image sensing means relative to an image sensing condition upon image sensing or an image sensing environment upon image sensing; and

extracting the predetermined object image from the sensed image by searching an image corresponding to the reference image changed in a shape and/or a size by said changing step;

wherein said image sensing parameter includes at least one of a focusing signal upon image sensing, focal length data of said image sensing means, and visual axis direction data and an illumination light condition upon image sensing.

2. The method according to claim 1, wherein the image corresponding to the reference image information is a pre-stored reference image and extracting the object image is executed by comparing the sensed image and the pre-stored reference image changed in an initial condition including an initial size, an initial position, or an initial color.

3. The method according to claim 1, wherein the image corresponding to the reference image information is an image under an image sensing condition, extracting the object image is executed by comparing a plurality of images under different image sensing conditions, and the reference image information is changed in correspondence with a relative relationship between the object and a background.

4. The method according to claim 1, wherein the reference image information is a template and the extracting object image is executed by comparing information of the sensed image and the template changed in a size.

5. An image information extraction apparatus comprising:

image sensing means for generating an image by sensing an object;

reference image storage means for storing a reference image to be compared with the image output from said image sensing means;

initialization means for automatically initializing the reference image by correcting a shape and/or a size of the reference image on the basis of at least an image sensing parameter of said image sensing means relative to an image sensing state when said image sensing means senses the object; and

image comparison means for comparing the reference image corrected in a shape and/or a size by said initialization means with the image output from said image sensing means, and determining the presence of the object when similarity between the two images exceeds a predetermined value;

wherein said image sensing parameter includes at least one of a focusing signal upon image sensing, focal length data of said image sensing means, and visual axis direction data and an illumination light condition upon image sensing.

6. The apparatus according to claim 5, further comprising image sensing mode extraction means for extracting said image sensing parameter.

7. The apparatus according to claim 5, wherein said initialization means sets an initial size of the reference image on the basis of an object distance upon image sensing or a focal length parameter of said image sensing means.

8. The apparatus according to claim 5, further comprising image sensing mode extraction means comprising visual axis direction data extraction means, and wherein said initialization means initializes a central position of the reference image in correspondence with a central position of an image generated by said image sensing means on the basis of a visual axis direction upon image sensing.

9. The apparatus according to claim 5, wherein said reference image storage means stores contour line data of a plurality of specific objects.

10. The apparatus according to claim 5, wherein said initialization means sets contour line data of the reference image as an initial contour, and sets a central position of the initial contour in the visual axis direction.

11. The apparatus according to claim 5, wherein said image comparison means uses an active contour as a contour model of the reference image, controls to vary positions of points on the contour on the basis of a predetermined evaluation function which assumes a minimum value when the function converges to a point on a contour line of the image generated by said image sensing means, thereafter, calculates similarity between a contour shape after convergence and a contour shape of the reference image or similarity between an image in a contour after convergence obtained by normalizing an image size so that the contour size becomes substantially equal to the reference image, and the reference image.

12. The apparatus according to claim 11, wherein said image comparison means outputs original image data in the contour after the active contour converges.

13. The apparatus according to claim 5, wherein said image comparison means outputs image data when the similarity between the reference image and the image generated by said image sensing means is larger than a predetermined threshold value.

14. An image information extraction apparatus comprising:

image sensing means for generating an image by sensing an object;

image sensing mode extraction means for extracting an image sensing mode when said image sensing means senses the object;

image sensing mode storage means for storing the image sensing mode extracted by said image sensing mode extraction means;

image storage means for storing the image output from said image sensing means;

reference image storage means for storing reference image information to be compared with the image output from said image sensing means;

initialization means for automatically initializing the reference image by correcting a shape and/or a size of the reference image on the basis of at least an image sensing parameter of said image sensing means relative to the image sensing mode extracted by said image sensing mode extraction means;

image comparison means for comparing the reference image corrected in a shape and/or a size by said initialization means with the image output from said image sensing means; and

image cutting means for cutting out an image on the basis of the output from said image comparison means;

wherein said image sensing parameter includes at least one of a focusing signal upon image sensing, focal length data of said image sensing means, and visual axis direction data and an illumination light condition upon image sensing.

15. The apparatus according to claim 14, wherein said image sensing mode extraction means extracts said image sensing parameter.

16. The apparatus according to claim 14, wherein said initialization means sets an initial size of the reference image on the basis of an object distance upon image sensing or a focal length parameter of said image sensing means.

17. The apparatus according to claim 14, wherein said image sensing mode extraction means comprises visual axis direction data extraction means, and said initialization means initializes a central position of the reference image in correspondence with a central position of an image generated by said image sensing means on the basis of a visual axis direction upon image sensing.

18. The apparatus according to claim 14, wherein said reference image storage means stores contour line data of a plurality of specific objects.

19. The apparatus according to claim 14, wherein said initialization means sets contour line data of the reference image as an initial contour, and sets a central position of the initial contour in the visual axis direction.

20. The apparatus according to claim 14, wherein said image comparison means uses an active contour as a contour model of the reference image, controls to vary positions of points on the contour on the basis of a predetermined evaluation function which assumes a minimum value when the function converges to a point on a contour line of the image generated by said image sensing means, thereafter, calculates similarity between a contour shape after convergence and a contour shape of the reference image or similarity between an image in a contour after convergence obtained by normalizing an image size so that the contour size becomes substantially equal to the reference image, and the reference image.

21. The apparatus according to claim 20, wherein said image comparison means outputs original image data in the contour after the active contour converges.

22. The apparatus according to 14, wherein said image comparison means outputs image data when the similarity between the reference image and the image generated by said image sensing means is larger than a predetermined threshold value.

23. An image information extraction method for searching for a specific object by executing:

image generation processing for generating an image by sensing an object;

extraction processing for extracting an image sensing state in the case that the image is generated by performing the image generation processing;

reference image reading processing for reading out reference image from storage means to be compared with the image generated by the image generation processing from storage means;

initialization processing for automatically initializing the reference image read out from the storage means by the reference image reading processing by correcting a shape and/or a size of the reference image on the basis of at least an image sensing parameter of an image sensing means relative to the image sensing state extracted by the extraction processing; and

image comparison processing for comparing the reference image corrected in a shape and/or a size by the initialization processing with the image generated by the image generation processing;

wherein said image sensing parameter includes at least one of a focusing signal upon image sensing, focal length data of said image sensing means, and visual axis direction data and an illumination light condition upon image sensing.

24. An image information extraction apparatus comprising:

image sensing means for generating an image by sensing an object;

image sensing mode storage means for storing an image sensing mode when said image sensing means senses the object;

image storage means for storing the image output from said image sensing means;

reference image storage means for storing a reference image to be compared with the image output from said image sensing means;

initialization means for automatically initializing the reference image by correcting a shape and/or a size of the reference image on the basis of at least an image sensing parameter of said image sensing means relative to the image sensing mode stored in said image sensing mode storage means; and

image comparison means for comparing the reference image corrected in a shape and/or a size by said initialization means with the image output from said image sensing means, and determining the presence of the object when similarity between the two images exceeds a predetermined value;

wherein said image sensing parameter includes at least one of a focusing signal upon image sensing, focal length data of said image sensing means, and visual axis direction data and an illumination light condition upon image sensing.

25. The apparatus according claim 24, further comprising image sensing mode extraction means for extracting said image sensing parameter.

26. The apparatus according to claim 24, wherein said initialization means sets an initial size of the reference image on the basis of an object distance upon image sensing or a focal length parameter of said image sensing means.

27. The apparatus according to claim 24, further comprising image sensing mode extraction means comprising visual axis direction data extraction means, and wherein said initialization means initializes a central position of the reference image in correspondence with a central position of an image generated by said image sensing means on the basis of a visual axis direction upon image sensing.

28. The apparatus according to claim 24, wherein said reference image storage means stores contour line data of a plurality of specific objects.

29. The apparatus according to claim 24, wherein said initialization means sets contour line data of the reference image as an initial contour, and sets a central position of the initial contour in the visual axis direction.

30. The apparatus according to claim 24, wherein said image comparison means uses an active contour as a contour model of the reference image, controls to vary positions of points on the contour on the basis of a predetermined evaluation function which assumes a minimum value when the function converges to a point on a contour line of the image generated by said image sensing means, thereafter, calculates similarity between a contour shape after convergence and a contour shape of the reference image or similarity between an image in a contour after convergence obtained by normalizing an image size so that the contour size becomes substantially equal to the reference image, and the reference image.

31. The apparatus according to claim 30, wherein said image comparison means outputs original image data in the contour after the active contour converges.

32. The apparatus according to claim 24, wherein said image comparison means outputs image data when the similarity between the reference image and the image generated by said image sensing means is larger than a predetermined threshold value.

33. A computer program product comprising a computer usable medium having computer readable program code means for extracting an object from a sensed image, including:

computer reading program code means for changing a shape and/or a size of reference image associated with object extraction on the basis of at least an image sensing parameter of an image sensing means relative to an image sensing condition upon image sensing or an image sensing environment upon image sensing; and

computer readable program code means for extracting the object from the sensed image by searching an image corresponding to the reference image;

wherein said image sensing parameter includes at least one of a focusing signal upon image sensing, focal length data of said image sensing means, and visual axis direction data and an illumination light condition upon image sensing.

34. A computer program product according to claim 33, said computer usable medium further comprising a database of images used for object extraction.