Method for extracting object images and method for detecting movements thereof

• US 5,751,831 A
• Filed: 06/05/1995
• Issued: 05/12/1998
• Est. Priority Date: 09/12/1991
• Status: Expired due to Fees

• Current Assignee: Fujifilm Corporation (Fujifilm Holdings Corporation)
• Original Assignee: Fuji Photo Film Co Limited (Fujifilm Holdings Corporation)
• Inventors: Ono, Shuji
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): Tran, Phuoc
• Application Number:
  US08/465,353
• Time in Patent Office: 1,072 Days
• Field of Search: 382/103, 382/107, 382/155, 382/156, 382/157, 382/158, 382/199, 382/236, 382/276, 348/154, 348/155, 348/169, 348/170, 348/171, 348/172, 364/516
• US Class Current: 382/103
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G06V 10/24 Aligning, centring, orientation detection or correction of the image
• CPC Class Codes: G06V 10/25 Determination of region of interest [ROI] or a volume of interest [VOI]
• CPC Class Codes: G06V 10/454 Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
• CPC Class Codes: G06V 10/46 Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features colour feature extraction G06V10/56
• CPC Class Codes: G06V 30/194 References adjustable by an adaptive method, e.g. learning
• CPC Class Codes: G06V 40/168 Feature extraction; Face representation

Abstract

In a method for extracting an object image, an extraction area for extraction of a candidate for a predetermined object image from an image is determined. The center point of a view window, which has a predetermined size, is caused to travel to the position of the candidate for the predetermined object image. The extraction area is determined in accordance with the size and/or the shape of the candidate for the predetermined object image, the center point of the view window being taken as a reference during the determination of the extraction area. The extraction of the candidate for the predetermined object image is carried out by using a neural network. Even if a plurality of object images, which are to be extracted, are embedded in a given image, the object images are extracted efficiently such that an object image, which has already been extracted, may not be extracted again.

Claims

1. A method for detecting a movement of an image, wherein a movement of an image is detected from the image having a background and an object, the method for detecting a movement of an image comprises the steps of:

i) cutting out a first image which falls in a region inside of a view window having a predetermined size which is smaller than the entire image, from the image,

ii) detecting a contour line of the object which is embedded in said cut-out first image,

iii) after a predetermined time has elapsed, cutting out a second image which falls in the region inside of said view window and which includes essentially the same object as said first image, from said image,

iv) detecting a contour line of the object which is embedded in said cut-out second image,

v) calculating a difference between said contour line which has been detected from said cut-out first image, and said contour line which has been detected from said cut-out second image, and

vi) automatically detecting a movement of the background of the image from said calculated difference.

2. A method for detecting a movement of an image as defined in claim 1, wherein a neural network is employed in order to carry out steps i) through vi).

3. A method for detecting a movement of an image as defined in claim 1, wherein said first and second images contain a candidate predetermined object image which is different from said background.

4. A method for detecting a movement of an image as defined in claim 1, further comprising the step of complex-log mapping the cut out first and second images into a complex-log plane and performing steps ii) and iv) based on the complex-log mapped first and second images.

5. A method for detecting a movement of an image, wherein a movement of an image is detected from the image having a background and an object, the method for detecting a movement of an image comprising the steps of:

i) cutting out a first image which falls in a region inside of a view window having a predetermined size which is smaller than the entire image, from the image,

ii) detecting a contour line of the object which is embedded in said cut-out first image,

iii) after a predetermined time has elapsed, cutting out a second image which falls in the region inside of said view window and which includes essentially the same object as said first image, from said image,

iv) detecting a contour line of the object which is embedded in said cut-out second image,

v) calculating a difference between said contour line which has been detected from said cut-out first image, and said contour line which has been detected from said cut-out second image,

vi) automatically detecting a movement of the background of the image from said calculated difference, and

vii) detecting a candidate for a predetermined object within the image by extracting the background of the image based on said movement of the background automatically detected in step vi).

6. A method for detecting a movement of an image as defined in claim 5, further comprising the step of complex-log mapping the cut out first and second images into a complex-log plane and performing steps ii) and iv) based on the complex-log mapped first and second images.

7. A method for detecting a movement of an image as defined in claim 5, further comprising the step of causing the view window to travel to the detected candidate for a predetermined object image after extracting the background of the image.

8. A method for detecting a movement of an image, wherein a movement of the image is detected from a plurality of images, in which identical object images are embedded together with backgrounds and which are taken at a plurality of times having a predetermined time difference therebetween, the method for detecting a movement of an image comprises the steps of:

i) (a) detecting contour lines of the object images, which are embedded in the plurality of images,

(b) calculating a first difference between the plurality of images based on said contour lines detected in step i) (a), and

(c) detecting a movement of the image in an in-plane parallel direction, the movement being detected from said first difference calculated in step i) b),

ii) (a) detecting contour lines of the object images, which are embedded in the plurality of images, said contour lines detected in step ii) (a) extending in a radial direction,

(b) calculating a second difference between the plurality of images based on said contour lines extending in the radial direction detected in step ii) (a), and

(c) detecting a movement of the image in an in-plane rotating direction, the movement being detected from said second difference calculated in step ii) (b),

iii) (a) detecting contour lines of the object images, which are embedded in the plurality of images, said contour lines detected in step iii) (a) extending in an annular direction,

(b) calculating a third difference between the plurality of images based on said contour lines extending in the annular direction detected in step iii) (a), and

(c) detecting a movement of the image in the radial direction, the movement being detected from said third difference calculated in step iii) (b), and

iv) detecting a movement of the image in accordance with said movement of said image in the in-plane parallel direction detected in step i), and at least one of said movements in the in-plane rotating direction detected in step ii) and said movement in the radial direction detected in step iii).

9. A method for detecting a movement of an image as defined in claim 8, wherein said steps ii) (a) and iii) (a) are carried out after the image has been transformed with a complex-log mapping into a complex-log mapped image by taking a viewpoint as a pole of mapping, and

the movements of the image detected in said steps ii) (c) and iii) (c) are detected from contour lines, which are detected from said complex-log mapped image.

10. A method for detecting a movement of an image as defined in claim 8 wherein a neural network is employed in order to carry out at least a single step selected from the group consisting of said step i) (c) of detecting the movement of the image in the in-plane parallel direction, said step ii) (c) of detecting the movement of the image in the in-plane rotating direction, said step iii) (c) of detecting the movement of the image in the radial direction, and said step iv) of detecting the movement of the image, wherein the movement of the image relates to movement of the background of the image.

11. A method for detecting a movement of an image as defined in claim 8 wherein components of a movement of the background of the image are compensated for in accordance with said detected movement of said image in the in-plane parallel direction, in the in-plane rotating direction, and/or in the radial direction, and a plurality of images, in which the components of the movement of the background have been compensated for, are formed.

12. A method for detecting a movement of an image as defined in claim 11 wherein the formation of the plurality of images, in which the components of the movement of the background have been compensated for, is carried out by using a neural network.