Method and apparatus for editing images using contour-extracting algorithm

• US 20070147700A1
• Filed: 07/25/2006
• Published: 06/28/2007
• Est. Priority Date: 12/28/2005
• Status: Abandoned Application

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Gangyu, Ma, Chen, Maolin, Jeong, Jin, Moon, Young, Kim, Ji
• Application Number:
  US11/491,968
• Publication Number:
  US 20070147700A1
• Time in Patent Office: Days
• US Class Current: 382/266
• CPC Class Codes: G06T 11/60 Editing figures and text; Combining figures or text
• CPC Class Codes: G06T 2207/20092 Interactive image processing based on input by user
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06T 7/194 involving foreground-background segmentation

Abstract

Disclosed herein is a method and apparatus for editing an image using an object contour extracted from an input image. The method of editing an image using a contour-extracting algorithm includes: inputting image data; extracting an object contour from the input image data; optimizing the extracted contour using the characteristics of the input image data; editing the input image data using the optimized contour; and outputting the edited image data.

Claims

1. A method of editing an image using a contour-extracting algorithm, the method comprising:

inputting image data;

extracting an object contour from the input image data;

optimizing the extracted contour using characteristics of the input image data;

editing the input image data using the optimized contour; and

outputting the edited image data.

2. The method of claim 1, wherein the extracting an object contour from the input image data further comprises:

extracting a position of an image object from the input image data;

extracting initial contour data from the input image data using a specific contour model;

extracting gradient information from the input image data; and

modifying the extracted initial contour data using the extracted gradient information;

wherein the image object may be an image of a person.

3. The method of claim 2, wherein the extracting a position of the image object further comprises detecting at least one of a face, eyes, and a skin color of the person from the input image data.

4. The method of claim 2, wherein the extracting initial contour data further comprises extracting a size of the person based on a distance between both eyes of the person, and subjecting the extracted person size to a model scaling to represent the object contour as control points.

5. The method of claim 2, wherein the extracting initial contour data further comprises representing the contour model by eigenvectors generated using training images.

6. The method of claim 2, wherein the extracting gradient information further comprises extracting gradient information from the input image data using a gradient vector flow (GVF).

7. The method of claim 2, wherein the modifying of the extracted initial contour data further comprises moving control points of the initial contour to a pixel having a high edge density.

8. The method of claim 7, wherein moving control points of the initial personal contour comprises moving the control points of the initial object contour using characteristics that alter a direction of a gradient information vector in a pixel having a high edge density.

9. The method of claim 1, wherein the optimizing the extracted contour further comprises:

retrieving the optimum contour by utilizing the input image data and a specific learned contour model; and

modifying the contour model by utilizing the retrieved optimum contour.

10. The method of claim 9, wherein the retrieving the optimum contour further comprises retrieving a control point of the optimum contour from current image data using the characteristics of the input image data and the specific learned contour model.

11. The method of claim 10, wherein the retrieving a control point of the optimum contour further comprises selecting a neighboring pixel in which a result value of an energy function for the input image data is a minimum and determining the selected neighboring pixel as a new control point.

12. The method of claim 11, wherein the energy function is an objective function which specifies a condition for deciding the control point corresponding to the objectcontour.

13. The method of claim 10, wherein the modifying the contour model further comprises adding a difference value between the retrieved control point and a control point of the contour model to the control point of the contour model.

14. The method of claim 1 further comprising:

receiving a request for correction of the contour from a user; and

adjusting a position of a control point for the extracted contour in response to the received request for correction of the contour,

wherein the optimizing the extracted object contour comprises optimizing the contour according to an energy function altered due to the adjusted position of the control point for the extracted contour.

15. The method of claim 1, wherein editing the input image data using the optimized contour comprises:

inserting the optimized contour into a predetermined background image; and

subjecting the inserted contour to an image matting.

16. The method of claim 15, wherein the inserting the optimized contour further comprises:

scaling the optimized contour to conform to a resolution of the predetermined background image; and

inserting the scaled contour into a desired position of the predetermined background image.

17. The method of claim 15, wherein the inserting the optimized contour further comprises:

receiving resolution information related to the predetermined background image, an object region and a person image;

calculating the scaling ratio between the person image and the background image based on the resolution information of the person image and the predetermined background image;

generating a bounding box using a largest width and height of the object region;

scaling the object region to conform to the calculated scaling ratio between the person image and the predetermined background image and a ratio of the bounding box; and

synthesizing the scaled object region with the predetermined background image.

18. The method of claim 15, wherein inserting the optimized contour further comprises designating a position where an edge density of the predetermined background image is lowest as a position for the object contour to be inserted, and inserting the optimized contour into the designated position.

19. The method of claim 1, wherein editing the input image data using the optimized contour further comprises:

detecting skin color of an image object from data about the optimized contour and segmenting a clothing region and a facial region of the object contour; and

adjusting a shape of the segmented clothing region and a brightness of the segmented facial region;

wherein the image object may be an image of a person.

20. The method of claim 1, wherein editing the input image data using the optimized contour further comprises adjusting a brightness of a background region of the input image data against the optimized contour.

21. A computer-readable recording medium storing therein a program to control an apparatus according to the method of claim 1.

22. An apparatus for editing an image using a personal contour-extracting algorithm, the apparatus comprising:

an image input section for inputting image data;

a contour-extracting section for extracting an object contour from the input image data applied to the contour-extracting section from the image input section;

a contour-optimizing section for optimizing the extracted contour applied to the contour-optimizing section from the contour-extracting section using characteristics of the input image data;

an image-editing section for editing the image data using the optimized extracted contour applied to the image-editing section from the contour-optimizing section; and

an image output section for outputting the edited image data applied to the image output section from the image-editing section.

23. The apparatus of claim 22, wherein the contour-extracting section detects at least one of a face, eyes, and a skin color of an image object from the input image data, and extracts initial contour data from the input image data using a specific contour model;

wherein the image object may be an image of a person.

24. The apparatus of claim 22, wherein the contour-optimizing section optimizes the extracted contour using characteristics of energy or an edge of the input image data.

25. The apparatus of claim 22, wherein the image-editing section synthesizes the optimized contour with a background image, edits at least one of a clothing shape or a face of the contour, or adjusts a brightness of a background region of the input image data.

26. A method of updating a contour model used to edit an image, comprising:

extracting an object contour from inputted image data;

optimizing the extracted contour using characteristics of the inputted image data and a specific learned contour model; and

updating the contour model to conform to the optimized object contour.

27. An apparatus for editing an image, comprising:

a contour-extracting section extracting an object contour from inputted image data;

a contour-optimizing section optimizing the extracted contour using characteristics of the inputted image data and a specific learned contour model; and

an image-editing section editing the image data using the optimized extracted contour.