METHOD AND SYSTEM OF ADAPTIVE REFORMATTING OF DIGITAL IMAGE

• US 20090180713A1
• Filed: 12/23/2008
• Published: 07/16/2009
• Est. Priority Date: 01/10/2008
• Status: Active Grant

• Current Assignee: S-Printing Solution Co., Ltd. (Samsung Electronics Co. Ltd.)
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: BUCHA, Victor, RYCHAGOV, Michael N., SAFONOV, Ilia
• Granted Patent: US 8,472,753 B2
• Time in Patent Office: Days
• US Class Current: 382/300
• CPC Class Codes: G06T 3/04 Context-preserving transformations, e.g. by using an importance map panospheric to cylindrical image transformations G06T3/12

Abstract

Method and system of adaptive reformatting of digital photos and document image. The method includes selecting a rule of calculation of function of value for a pixel depending on results of analyzing an image, generating a map of value and a map of history, changing a size of the image by adding/removing at least one found path by means of interpolation, and repeating generation of the maps of values and history and change of a size of the image until a demanded size of the image is received or a sum of path values exceeds a predetermined threshold.

Claims

1. A system of adaptive reformatting of digital photos, which comprises a module to select a rule of calculation of function of value for a pixel, a generator of maps of values and history, a module to analyze the map of values and select a path, a module to add a path to an image, a module to remove a path from an image, a scaling module, a module to weaken the map of history, and a module to update the map of values,

wherein an output of the module to select a rule of calculation of function of values for a pixel is connected with an input of the generator of maps of values and history, and also with an input of the module to update the map of values, an output of the generator of maps of values and history is connected with an input of the module to analyze the map of values and select a path, an output of the module to analyze the map of values and select a path is connected with inputs of the module to add a path to the image, of the module to remove a path from the image, and of the scaling module, outputs of the module to add a path to the image and of the module to remove a path from the image are connected with an input of the module to weaken the map of history, an output of the module to weaken the map of history is connected with an input of the module to update the map of values, and an output of the module to update the map of values is connected with an input of the module to analyze the map of values and select a path,

wherein the module to select a rule of calculation of function of value for a pixel selects a rule of calculation of function of value for each pixel from a set of predetermined rules depending on information contained in the image,

wherein the generator of maps of values and history generates an initial map of values and an initial map of history, using the selected rule,

wherein the module to analyze the map of values and select a path determines at least one horizontal path from connected pixels, where an initial pixel of a path is on the left border of the image, a final pixel of the path is on the right border of the image, and the quantity of pixels in the path is equal to the width of the image of pixels, and also determines at least one vertical path from connected pixels, where an initial pixel of a path is on the top border of the image, a final pixel of the path is on the bottom border of the image, and the quantity of pixels in a path is equal to the height of the image of pixels, and if a sum of pixel values of all found paths is greater than or equal to a predetermined threshold, the image is transferred to the scaling module, and if the sum is less than the predetermined value, the image and the maps of values and history are transferred to the module to add a path to the image or the module to remove a path from the image depending on a current and required ratio of the sides of the image,

wherein the module to add a path to the image increases the current image by adding the least one found path to the image using interpolation for several nearest pixels,

wherein the module to remove a path from the image reduces the current image by removing the least one found path from the image,

wherein the module to weaken the map of history reduces the values in the map of history, which correspond to environment of pixels of paths removed or added earlier,

wherein the module to update the map of value recalculates a map of values, using the previously selected rule and map of history.

20. An image forming apparatus using the system claimed in claim 1.

2. A method of adaptive reformatting of digital photos and document images, the method comprising:

selecting a rule of calculation of function of value for a pixel depending on results of analyzing an image;

generating a map of value and a map of history;

changing a size of the image by adding/removing at least one found path by means of interpolation, wherein both of horizontal and vertical paths are used, and the horizontal path is a chain of the connected pixels where an initial pixel of the horizontal path is on the left border of the image, a final pixel of the horizontal path is on the right border of the image, and the quantity of pixels in the horizontal path is equal to the width of the image of pixels, whereas the vertical path is a chain of the connected pixels where an initial pixel of the vertical path is on the top border of the image, a final pixel of the vertical path is on the bottom border of the image, and the quantity of pixels in the vertical path is equal to the height of the image of pixels, and search for the path is executed by means of minimization of a sum of pixel values of the path; and

repeating generation of the maps of values and history and change of a size of the image until a demanded size of the image is achieved or the sum of path values exceeds a predetermined threshold.

3. The method of claim 2, wherein detecting a face image of people is used when the rule of calculation of function of value is selected.

4. The method of claim 2, wherein a detector of city stages and architectural constructions is used when the rule of calculation of function of value is selected.

5. The method of claim 2, wherein a high-frequency filter is used when the map of value is generated.

6. The method of claim 2, wherein a function describing colors typical for human skin is used when the map of value is generated.

7. The method of claim 2, wherein the changing a size of the image increases the image by adding at least one path to the image, comprising:

defining a path with a minimum sum of values;

adding new pixels next to pixels of the path using interpolation by several nearest pixels;

weakening values in the map of history and updating the map of history; and

updating values in the map of values.

8. The method of claim 2, wherein the changing a size of the image reduces the image by removing at least one path from the image, comprising:

defining a path with a minimum sum of values;

removing pixels of the path and correcting values of the pixels, thereby bordering with the pixels of the removed path;

weakening values in the map of history and updating the map of history; and

updating values in the map of value.

9. The method of claim 2, wherein the weakening values in the map of history uses the following equation:

H(i,j)=H(i,j)<sup>−

1</sup>−

C,

where H(i,j) denotes a value of element of the map of history, H(i,j)<sup>−

i</sup> denotes a value of element of the map of history before weakening, and C denotes a predetermined constant value.

21. An image forming apparatus using the method claimed in claim 2.

10. An image forming apparatus comprising:

a module to select a rule of calculation of function of value for a pixel from a set of predetermined rules depending on information contained in an image;

a generator to generate an initial map of values and an initial map of history, using the selected rule;

a module to search at least one horizontal or vertical path with a minimum sum of values by analyzing the initial map of values; and

a module to change a size of the image by adding or removing a path to or from the image using the searched path.

11. The image forming apparatus of claim 10,

wherein the horizontal path is a chain of pixels where an initial pixel is on the left border of the image, a final pixel is on the right border of the image, and the quantity of pixels is equal to the width of the image of pixels,

wherein the vertical path is a chain of pixels where an initial pixel is on the top border of the image, a final pixel is on the bottom border of the image, and the quantity of pixels is equal to the height of the image of pixels.

12. The image forming apparatus of claim 10, further comprising:

a module to scale the image if a sum of pixel values of the paths is greater than or equal to a predetermined threshold.

13. The image forming apparatus of claim 10, wherein the module to change a size of the image comprises:

a module to add the searched path to the image; and

a module to remove the searched path from the image.

14. The image forming apparatus of claim 13, wherein the module to add the searched path adds at least one path by adding new pixels next to pixels of the searched path using interpolation by several nearest pixels, by weakening values in the map of history and updating the map of history, and updating values in the map of values.

15. The image forming apparatus of claim 13, wherein the module to remove the searched path removes at least one path by defining a path with a minimum sum of values, removing pixels of the path and correcting values of the pixels, thereby bordering with the pixels of the removed path, weakening values in the map of history and updating the map of history, and updating values in the map of value.

16. The image forming apparatus of claim 10, wherein the module to select a rule of calculation of function of value uses detecting a face image of people.

17. The image forming apparatus of claim 10, wherein the module to select a rule of calculation of function of value uses detecting city stages and architectural constructions.

18. The image forming apparatus of claim 10, wherein the generator uses a high frequency filter.

19. The image forming apparatus of claim 10, wherein the generator uses a function describing colors typical for human skin.