KR101483611B1 - Method and Terminal for Extracting a Object from Image - Google Patents

Abstract

The present invention relates to a method for extracting an object from an image.
A method for extracting an object from an image according to the present invention includes: a first step of extracting one or more unit objects based on a degree of similarity between pixels in an unit object extracting unit; A unit object extraction unit displaying the unit object through a display unit; A third step of determining whether two or more unit objects are selected through the input unit among the unit objects displayed on the display unit of the object setting unit; And a fourth step of the object setting unit combining the unit objects selected in the third step and determining them as objects.

Description

[0001] The present invention relates to a method and a terminal for extracting an object from an image,

The present invention relates to a method and apparatus for extracting an object from an image, and a method and apparatus for creating a new object by grouping a plurality of objects.

In modern society, many digital contents are accumulated and stored by information.

One of the biggest advantages of digital content, including text / audio / video content, is that it is easier for users to rework in the computer environment. For example, a document text file can be easily edited using a text editor. Similarly, the image content can also be edited to edit some or all of the pixels that make up the image. These advantages can make digital content more widely available.

On the other hand, in the process of reprocessing and transmitting a large number of contents, the original content of the content becomes unclear and the title of the content is also changed. Particularly, since the image content is not easy to search using a 'keyword' like a text file, it is difficult to search image content.

In recent years, there have been proposed ways to increase the use and function of the Internet search based on image content search, but the efficiency is still small. Particularly, there are currently no measures to identify objects for image retrieval.

For example, in order to search for text, there is a method of searching through a combination of keywords in addition to one completed keyword, but the searching method is very limited in the process of searching for image contents.

Accordingly, it is an object of the present invention to provide a method and apparatus for extracting an object for efficiency of an image retrieval method.

According to an aspect of the present invention, there is provided a method of extracting objects from an image, the method comprising: a first step of extracting one or more unit objects based on a degree of similarity between pixels in an image; A unit object extraction unit displaying the unit object through a display unit; A third step of determining whether two or more unit objects are selected through the input unit among the unit objects displayed on the display unit of the object setting unit; And a fourth step of the object setting unit combining the unit objects selected in the third step and determining them as objects.

In one embodiment, the input unit uses a touch panel coupled to the display unit, and in the third step, the object setting unit can determine whether unit objects exist in the drag region on the touch panel.

In the third step, when the object setting unit has one or more touch operations on the touch panel, it can be determined whether or not the unit objects exist at the touch point.

If there is no selection operation of two or more unit objects in the third step, the object setting unit may determine each unit object as an object.

Or setting the new object by grouping the object and the unit object when the object and the unit objects are selected from the user after the object is determined in the third step.

In addition, after the object is determined in the third step, when there is a selection of objects from the user, the object setting unit may group the objects to set the object pair.

A terminal having a function of extracting an object from an image according to the present invention includes a unit object extracting unit for extracting one or more unit objects based on similarity between pixels in an image and displaying the unit object through a display unit; An input unit for selecting one or more unit objects to be displayed on the display unit, the touch panel being connected to the display unit or the mouse unit for a computer; And an object setting unit for determining unit objects selected through the input unit as objects.

According to the method and apparatus for extracting an object according to the present invention, a unit object is extracted based on the similarity of pixel information in an image, and an object is created based on a unit object selected through a user interface. Therefore, according to the method and apparatus for extracting an object according to the present invention, it is possible to easily and efficiently create an object for image search.

In addition, according to the method and apparatus for extracting an object according to the present invention, a new object can be created by combining unit object-unit object, unit object-object, and object-object through a simple user interface. That is, according to the method and apparatus for extracting an object according to the present invention, it is possible to extract an object of an image for efficient and accurate searching through combining of images similar to a combination of search terms of text.

1 is a block diagram illustrating a configuration of a terminal having an object extraction function according to the present invention;
2 is a flowchart showing an object extraction method according to the present invention;
3 is a schematic diagram showing an example of a method of displaying a unit object;
FIGS. 4-6 illustrate an embodiment of a method for selecting a unit object. FIG.
Figure 7 illustrates a method of grouping unit objects and objects.
8 is a diagram illustrating a method of grouping objects and objects.
9 is a flowchart showing an embodiment of a method of extracting a unit object;
10 is a schematic diagram showing an example of a method of displaying a unit object;

Hereinafter, embodiments of the present invention in which the above object can be specifically realized will be described with reference to the accompanying drawings. In the description of the embodiments, the same names and symbols are used for the same components, and further description thereof will be omitted.

1 is a block diagram showing the configuration of a terminal 100 having an object extraction function according to the present invention.

The terminal 100 may be a portable terminal, such as a smart phone or a tablet PC, in which an application program can be installed in the form of an application.

The memory 130 is for storing images in a digital data format.

The display unit 130 may be a means for providing a user interface for displaying an image, in particular, for selecting a unit object from an image. That is, the display unit 130 is coupled with the touch panel 132 including the touch sensor, so that the pixels of the image displayed through the display unit 130 can be selected through the touch panel 132.

The controller 140 controls the overall operation of the terminal 100, and particularly extracts unit objects and groups the selected unit objects to generate objects.

The control unit 140 may be implemented as a processor chip, which is basically a hardware possessed by the terminal 100, and a program operated by the chip. That is, the overall functions of the terminal 100 include an operating system (OS) program executed by the processor chip, an application program installed on the OS program and controlling the hardware configuration of the terminal 100, Lt; RTI ID = 0.0 > a < / RTI > For example, a part of the configuration including the object extraction unit 142 of the control unit 140 may be installed in the memory 120 as an application form.

Hereinafter, the object extracting unit 142, which is a constitution of the control unit 140 centering on the implementation of the present invention, will be described in this specification.

The object extracting unit 142 extracts a foreground distinguished from the background in the image as an object. For this, the object extracting unit 142 includes a unit object extracting unit 144 and an object setting unit 146.

The unit object extracting unit 144 groups the pixels having strong similarity in the original image stored in the memory 130 into unit objects and displays the unit objects through the display unit 130. At this time, the unit object extracting unit 144 may extract the unit object based on the color similarity considering the distance between the pixels. A detailed method of extracting the unit object by the unit object extracting unit 144 will be described later.

The object setting unit 146 groups the unit objects selected by the user among the unit objects displayed on the display unit 130 and determines the objects as objects. In addition, the object setting unit 146 may group the unit object and the object, and may group the object and the object.

A method of extracting an object from an original image stored in the memory 130 using the object extracting unit 142 will be described with reference to the flowchart of FIG.

<Unit object extraction and display: S201>

The unit object extraction unit 144 extracts a unit object from an image stored in the memory 120 or an original image transmitted from a search server (not shown). A unit object is a partial image in which pixels having a relevance in an image are combined, and can be determined by combining pixels based on the similarity of hues.

3, the unit object extraction unit 144 displays the extracted first through third unit objects 401, 402, and 403 through the display unit 130. The unit object extraction unit 144 extracts the first, second, The outermost pixels of the unit objects 401, 402, and 403 can be represented by the edges of the first, second, and third unit objects 401, 402, and 403 unit objects 401,

<Set selected unit objects as objects: S203, S205>

In step S201, the object setting unit 146 combines unit objects selected from among the first through third unit objects 401, 402, and 403 displayed on the display unit 130 and sets them as objects.

For this, the object extracting unit 142 recognizes that the touch operation is performed on the touch panel 132, thereby confirming the unit object selected by the user. That is, the object setting unit 146 determines whether the pixel at the contact point is included in the first through third unit objects 401, 402, and 403 in the touch panel 132, (401, 402, 403), the unit object of the contact point is determined as a part of the object. For example, in the example shown in FIG. 4, when the contact operation is performed on the point a, the object setting unit 146 regards the first unit object 401 as a unit object included in the object. When the touch operation is performed on a pixel other than the area included in the first through third unit objects 401, 402, and 403, the object setting unit 146 ignores the touch operation. Also, the object setting unit 146 determines whether the pixels corresponding to each touch operation belong to the unit object when not only the touch operation on one point but also the consecutive pixels are performed by the drag operation.

At this time, the object setting unit 146 combines a plurality of unit objects selected when two or more unit objects are selected, and determines the object as an object. An embodiment for grouping a plurality of unit objects as described above will be described below.

* First Embodiment - Grouping of Unit Objects by Drag Operation

As shown in FIG. 4, when the touch operation is performed on the point a, the object setting unit 146 regards the first unit object 401 as a unit object included in the combination of objects.

5, when the drag operation is performed from the point a to the point c, the object setting unit 146 determines whether or not there is a unit object including the pixels contacted by the drag operation, The unit objects are grouped and determined as the object 410. 5, the object setting unit 146 sets the first object 410 as a combination of the first through third unit objects 401, 402, and 403.

If the pixel selected in the drag operation is not included in the first through third unit objects 401, 402, and 403, the object setting unit 146 ignores the corresponding pixel region.

* Second Embodiment - Grouping of Unit Objects by Click Operation

6 is a diagram illustrating a second embodiment for selecting a unit object. The object setting unit 146 may select one or more of the first through third unit objects 401, 402, and 403 based on an operation clicked on the touch panel 132, as shown in FIG. That is, when an individual clicking operation is performed on the point a, point b and point c, the object setting unit 146 sets the first through third unit objects 401, 402, and 403 including the pixels of the point a, point b, The first object 410 is set.

Third Embodiment - Grouping Objects and Unit Objects

The object setting unit 146 that sets the object based on the unit object selected by the first and second embodiments can reset the object by combining the selected unit objects.

In a state where the first object 410 is determined by grouping the first through third unit objects 401, 402, and 403 using the above-described embodiment, when an additional unit object is selected, You can set a new object that contains the object.

For example, as shown in FIG. 7, when the first object 410 and the fourth unit object 404 are selected by performing the drag operation from the point ⓔ to the point 과 as in the method of the first embodiment, the object setting unit 146 And sets a second object 420 in which the first object 410 and the fourth unit object 404 are grouped.

At this time, the object setting unit 146 may set the second object 420 as in the second embodiment for the operation of clicking the first object 410 and the fourth unit object 404 instead of the drag operation .

* Fourth Embodiment - Object Pair Setting by Grouping Objects and Objects

The object setting unit 146 may set the object pairs by grouping the objects set through the first to third embodiments.

8, when the first object 410 and the third object 430 are selected by the drag operation, the object setting unit 146 sets the first object 410 and the third object 430 as And sets the object pair 450 that contains it.

At this time, the object setting unit 146 may set the object pair 450 by clicking the first object 410 and the third object 430 instead of the drag operation.

Also, when a pixel other than an object or a unit object is selected in the drag operation process, the object setting unit 146 ignores the corresponding pixel area.

Such an object pair 450 may be a group of objects of the same type to select a plurality of objects as a single pair of objects, or a group of objects of different types may be grouped to select a pair of objects as a pair of objects. That is, the object pair 450 can be used as a condition for searching an image that includes a plurality of objects, or a pair of objects such as a mug and a pedestal or a spoon and a chopstick.

If a plurality of unit objects are not selected in step S203, the object setting unit 146 may determine the first through third unit objects 401, 402, and 403 as objects.

As described above, the object extracting unit 142 according to the present invention can generate objects by grouping unit objects.

In this case, the first through third unit objects 401, 402, and 403 can be extracted based on the degree of similarity in the original image. For example, a method of extracting a unit object based on the similarity of colors will be described with reference to FIG. Respectively.

<Calculation of Bonding Force between Pixels: S901>

In order to extract the object from the original image, the unit object extracting unit 144 first divides the image based on the color similarity. At this time, the unit object extracting unit 144 divides the original image based on the 'binding force' calculated based on the color similarity according to the distance between the pixels. The binding force is a relative value with respect to the degree of coupling between two contrasting pixels, and is defined as the following Equation (1).

Here, 'color similarity' represents a degree of similarity between the color values of two pixels included in the pair of contrast pixels. 'Interpixel distance' represents the relative distance between two pixels included in a pixel pair. That is, the binding force is a value indicating the degree of coupling of pixel pairs having a larger value as the two pixels are similar in color and the distance between the two pixels is closer.

In order to obtain such a binding force, the unit object extracting unit 144 selects a pixel pair consisting of two pixels in the original image. The unit object extracting unit 144 may select all pairs of pixels that can be combined, or may select a pair of pixels only for pixels located within a certain distance.

&Lt; &lt; S902 &gt; Determining a unit object by combining pixels having a bonding strength equal to or greater than a reference value:

Based on the binding force calculated in step S902, the unit object extracting unit 144 calculates a binding force for each pixel pair, which is a combination of two pixels. If the calculated binding force is equal to or greater than a predetermined reference value, do. At this time, the reference value can be obtained experimentally.

Through this process, the unit object extracting unit 144 can extract the unit object by excluding the pixel pairs in which the colors are largely different or the distances between the pixels are far apart, and by connecting pixel pairs having strong binding force. That is, as shown in FIG. 10, the unit object extracting unit 144 can set a unit object by combining the original image with pixels of P1 to P7, pixels of P8 to P13, pixels of P14 to P18, and pixels of P19 to P20 have.

At this time, the unit object extracting unit 144 may not display the corresponding unit object on the display unit 130 when the number of pixels included in the unit object is smaller than a predetermined threshold, even though the unit object extracting unit 144 is strong. For example, the pixels P19 to P20 may not be displayed as a unit object on the display unit 130 even if color similarity is strong.

In the above-described embodiment, the terminal 100 has been described on the assumption that a portable terminal such as a smart phone is assumed, and a touch panel 132 is coupled to the display unit 130. [ However, it is apparent that the embodiment of the present invention can also be used as a terminal such as another personal computer. In the case of using a personal computer, the object extracting unit 142 may be installed in a computer as an application program, and the touch panel 132 for receiving pixels from the image displayed on the display unit 130 may be a mouse device for a computer .

In addition, the present invention is not limited to the terminal, and the object extraction unit 142 for object extraction may be included in the configuration of the server having the search function. In this case, the server may extract the unit object from the image stored in the terminal 100 or the image stored in the database of the server, and display the extracted unit object through the display unit 130 of the terminal 100. The terminal 100 may select a pixel using the mouse device or the touch panel 132, and the server may determine whether the pixel selected by the terminal corresponds to a unit object and set the object.

It is to be understood by those skilled in the art that the present invention may be embodied in many other forms without departing from the spirit and scope of the invention, Accordingly, the above-described embodiments are to be considered illustrative and not restrictive, and all embodiments falling within the scope of the appended claims and their equivalents are intended to be included within the scope of the present invention.

Claims (7)

A method of extracting an object from an image,
A first step of extracting at least one unit object based on the degree of similarity between pixels in the unit object extracting unit;
A second step of the unit object extracting unit displaying the unit object through a display unit;
A third step of determining whether two or more unit objects are selected through the input unit among the unit objects displayed on the display unit by the object setting unit;
A fourth step of the object setting unit combining the unit objects selected in the third step and determining the object as an object; And
And a fifth step of the object setting unit determining each of the unit objects as an object when there is no selection operation of two or more unit objects in the third step.
The method according to claim 1,
Wherein the input unit uses a touch panel coupled to a display unit,
Wherein the object setting unit determines whether the unit objects exist in the drag area on the touch panel.
3. The method of claim 2,
Wherein the third step determines whether the unit objects exist at a contact point when the object setting unit has at least one contact action on the touch panel.
delete 4. The method according to any one of claims 1 to 3,
If there is a selection of the object and the unit objects from the user after determining the object in the third step,
Wherein the object setting unit groups the object and the unit object to set a new object.
4. The method according to any one of claims 1 to 3,
If there is a selection of the objects from the user after determining the object in the third step,
Wherein the object setting unit groups the objects and sets an object pair.
A terminal having a function of extracting an object from an image,
A unit object extraction unit extracting one or more unit objects based on similarity between pixels in an image and displaying the unit objects through a display unit;
An input unit for selecting at least one unit object displayed on the display unit using a touch panel or a computer mouse device coupled to the display unit; And
When two or more unit objects are selected through the input unit among the unit objects displayed on the display unit, the selected unit objects are combined to determine an object, and when there is no selection operation of two or more unit objects through the input unit And an object setting unit for determining the unit objects in the image as an object.

Images