KR20220075686A - Method for detecting mosaic image - Google Patents

Abstract

A mosaic image detection method according to an embodiment of the present invention is a method of detecting a mosaic in an image using a mosaic detection system, the method comprising: selecting a target image that is at least one of a plurality of images constituting the image; selecting a preset target object on the target image; generating a converted target image by converting the target object to a preset resolution lower than the resolution of the target image; calculating an anchor point of the target object on the target transformation image; calculating vertical and horizontal edges of the target object based on the calculated anchor point of the target object; and determining whether the target object is a mosaic by using the calculated horizontal and vertical edges of the target object.

Description

Mosaic image detection method

The present invention relates to a mosaic image detection method, and more particularly, to a mosaic image detection method for detecting whether there is a mosaic in an image using a mosaic detection system.

In recent years, illegal videos have been on the rise.

In the eight months of 2019, the number of illegal reproductions detected on YouTube reached 9,000, the same as last year.

According to data from the Korea Copyright Protection Agency, as a result of monitoring major websites, a total of 8,833 illegal reproductions were discovered on YouTube from early 2019 to August 14, 2019.

This is close to the 8,880 cases detected in 2018.

When the Copyright Protection Review Committee of the Korea Copyright Protection Agency finds that illegal reproductions have been transmitted, it goes through deliberation and recommends to the online service provider a warning about the copy/transmitter and the deletion and suspension of transmission of posts, but YouTube does not comply with the Korean Copyright Act. There has been no case of recommending corrective measures due to difficulties in administrative measures.

Accordingly, it is pointed out that it is necessary to prepare a way to take administrative measures on overseas sites such as YouTube to prevent copyright infringement of domestic content.

On the other hand, the illegal video material may be divided into a mosaic-processed image and a non-mosaic image.

In judging illegal video, the manager must first examine whether the video is a mosaic-processed video.

However, there is a hassle in that the manager has to check the images one by one and determine whether the mosaic has been processed.

Meanwhile, Korean Patent Registration No. 10-1903523 (September 21, 2018) discloses a method for detecting a mosaic based on edge projection in video data.

However, since the edge is detected directly from the input image, if the image quality of the input image is not good, there is a strong possibility of determining it as a mosaic.

The present invention is to solve the above problems, and to provide a mosaic image detection method capable of more accurately detecting a mosaic-processed image even in a low-quality image.

The problem to be solved by the present invention is not limited to the above-mentioned problems, and the problems not mentioned will be clearly understood by those of ordinary skill in the art to which the present invention belongs from the present specification and the accompanying drawings. .

A mosaic image detection method according to an embodiment of the present invention is a method of detecting a mosaic in an image using a mosaic detection system, the method comprising: selecting a target image that is at least one of a plurality of images constituting the image; selecting a preset target object on the target image; generating a converted target image by converting the target object to a preset resolution lower than the resolution of the target image; calculating an anchor point of the target object on the target transformation image; calculating vertical and horizontal edges of the target object based on the calculated anchor point of the target object; and determining whether the target object is a mosaic by using the calculated horizontal and vertical edges of the target object.

According to the method for detecting a mosaic image according to an embodiment of the present invention, there is an advantage that a mosaic-processed image can be more accurately detected even in a low-quality image.

Effects of the present invention are not limited to the above-described effects, and effects not mentioned will be clearly understood by those of ordinary skill in the art to which the present invention pertains from the present specification and accompanying drawings.

1 is a schematic block diagram of a mosaic image detection system implementing a mosaic image detection method according to an embodiment of the present invention;
2 is a schematic flowchart of a mosaic image detection method according to an embodiment of the present invention.
3 is an exemplary diagram for explaining mosaic in a low-resolution image.
4 to 6 are exemplary images for explaining a mosaic image detection method according to an embodiment of the present invention.

Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings. However, the spirit of the present invention is not limited to the presented embodiments, and those skilled in the art who understand the spirit of the present invention may add, change, delete, etc. other elements within the scope of the same spirit, and may use other degenerative inventions or the present invention. Other embodiments included within the scope of the invention may be easily proposed, but this will also be included within the scope of the invention.

A mosaic image detection method according to an embodiment of the present invention is a method of detecting a mosaic in an image using a mosaic detection system, the method comprising: selecting a target image that is at least one of a plurality of images constituting the image; selecting a preset target object on the target image; generating a converted target image by converting the target object to a preset resolution lower than the resolution of the target image; calculating an anchor point of the target object on the target transformation image; calculating vertical and horizontal edges of the target object based on the calculated anchor point of the target object; and determining whether the target object is a mosaic by using the calculated horizontal and vertical edges of the target object.

In addition, the calculating of the anchor point of the target object may include calculating the anchor point of the target object by converting the target transformed image to a resolution of the target image.

Also, in the calculating of the anchor point of the target object, the anchor point of the target object may be defined as a center point of a region having a boundary in which the contrast is greater than or equal to a preset reference on the target transformation image.

In addition, the calculating of the vertical and horizontal edges of the target object may include calculating a boundary in which the contrast is different than a preset reference based on the anchor point of the target object, and the boundary is based on the anchor point of the target object A distance greater than or equal to a preset distance may be defined as vertical and horizontal edges of the target object.

In addition, the determining whether the target object is a mosaic may include defining the target image as a mosaic when the calculated vertical and horizontal edges of the target object are greater than or equal to a preset number.

In addition, selecting a comparison area other than the target object on the target image; generating a comparison conversion image by converting the comparison area to a preset resolution lower than the resolution of the target image; calculating an anchor point of the comparison region on the comparison transformed image; Calculating vertical and horizontal edges of the comparison region based on the calculated anchor point of the comparison region; further comprising, determining whether the target object is a mosaic, the calculated horizontal and horizontal edges of the comparison region Whether the target image is a mosaic may be determined by comparing a vertical edge with horizontal and vertical edges of the target object.

In addition, the selecting of the comparison area may include selecting the comparison area having the same size as an area occupied by the target object on the target image.

In addition, the selecting of the comparison region includes generating a plurality of divided regions having the same size with respect to at least a portion of the region excluding the region occupied by the target object on the target image, and selecting a plurality of the divided regions to The comparison area having the same size as the area occupied by the target object may be selected.

Elements having the same function within the scope of the same idea shown in the drawings of each embodiment will be described using the same reference numerals.

1 is a schematic block diagram of a mosaic image detection system implementing a mosaic image detection method according to an embodiment of the present invention.

2 is a schematic flowchart of a method for detecting a mosaic image according to an embodiment of the present invention.

3 is an exemplary diagram for explaining mosaic in a low-quality image.

4 to 6 are exemplary images for explaining a method for detecting a mosaic image according to an embodiment of the present invention.

In the accompanying drawings, in order to more clearly express the technical spirit of the present invention, parts that are not related to the technical spirit of the present invention or that can be easily derived from those skilled in the art have been simplified or omitted.

1 to 6 , the mosaic image detection method according to an embodiment of the present invention is implemented by the mosaic detection system 10, which means a method of detecting whether a mosaic has been processed in an image. can

In general, a mosaic detection method determines whether a mosaic is a mosaic by detecting horizontal and vertical edges in an image constituting an image.

In general, the mosaic-treated area has the following characteristics.

First, the colors of pixels located in the mosaic block are the same. Therefore, there is no edge inside the mosaic block.

Second, the sizes of the mosaic blocks are the same.

Third, the mosaic blocks are adjacent to each other and form a cluster.

Fourth, there is no edge between adjacent mosaic blocks having the same intensity value.

Here, Fig. 3 (a) is an enlarged view of a mosaic-processed image, and Fig. 3 (b) is an enlarged view of an image of a low-quality image or a portion where mosaic is not processed. When comparing both, both horizontal and vertical edges are There is a problem in that it has, that is, FIG. 3(b) is not a substantially mosaic-processed image, but can be detected as a mosaic-processed image.

The present invention is intended to solve the above problems, and may be implemented by the mosaic detection system 10 .

1 is a schematic block diagram of components constituting the mosaic detection system 10 .

As shown in FIG. 1 , the mosaic detection system 10 includes a communication unit 300 that receives an image to be determined from an external electronic device P, and a memory that stores data/information/arithmetic expressions necessary for mosaic detection. Unit 200, an input unit 400 that allows an administrator to input data/information/arithmetic expressions required for mosaic detection, and an output unit 500 that displays mosaic detection results and data/information/arithmetic expressions required for mosaic detection and a control unit 100 that implements mosaic detection.

The control unit 100 may be configured to calculate predetermined data/information and determine whether to perform mosaic processing in order to implement a mosaic detection method to be described below.

That is, the communication unit 300 may obtain the determination target image from the external electronic device P, and the control unit 100 may implement the mosaic detection method to be described below.

Hereinafter, a mosaic detection method according to an embodiment of the present invention implemented by the controller 100 will be described in more detail with reference to FIGS. 2 to 6 .

The mosaic detection method may include selecting a target image that is at least one of a plurality of images constituting the image of the judgment target ( S10 ).

In the step (S10) of selecting the target image, in the image received through the communication unit 300, the target image (A), which is a judgment target for determining whether a mosaic has been arbitrarily processed among a plurality of images constituting the image You can choose.

For example, in the step of selecting the target image ( S10 ), the target image A may be generated by selecting one or at least two or more from among a plurality of images constituting the image.

Here, as an example, the mosaic detection method may further include selecting a preset target object on the target image A ( S20 ).

To explain this in more detail, the step (S20) of selecting the target object does not determine whether mosaic processing has been performed by calculating the entire target image (A) as a target, but on the target image (A) as desired by the administrator. It may be a step that exists to determine whether the part has been mosaicized.

For example, if the administrator wants to know whether a part defined as a face on the target image (A) has been mosaicized, a command to extract a part for the face as the target object (B) can be given. Accordingly, the step (S20) of selecting the target object for selecting the face part as the target object (B) may be implemented.

Since extracting a region for a specific part, such as a face, from an image is a known technique, a detailed description will be omitted.

On the other hand, as shown in Fig. 4, the step of selecting the target object (S20) is to select the target object (B) for the face region by a preset command to extract the face from the target image (A). You can extract/select.

Here, as an example, the method for detecting a mosaic image may further include generating a converted target image by converting the target object B to a preset resolution lower than the resolution of the target image A ( S30 ). .

The step of generating the target transformation image ( S30 ) may refer to a step of pre-processing the target object B in order to more accurately find the anchor point P, which will be described below.

To explain this in more detail, as shown in FIG. 4 , the step of generating the target transformed image ( S30 ) is to reduce the resolution of the target object B having the same resolution as the target image A to a relatively low low resolution. It is possible to generate the target transformed image (C) for the target object (B) by converting and/or black-and-white processing.

As a result, the object conversion image C may be simplified in contrast, chromaticity, and the like.

Here, as an example, the method for detecting a mosaic image may further include calculating an anchor point of the target object on the target transformed image C ( S40 ).

The anchor point P may refer to a reference point for determining whether to perform mosaic processing.

For example, the anchor point P may be a reference point for finding vertical and horizontal edges in the step S50 of calculating vertical and horizontal edges of a target object, which will be described below.

On the other hand, as shown in FIG. 4 , in the step of calculating the anchor point of the target object ( S40 ), the anchor point P can be generated on a portion with relatively low brightness (white portion) on the converted image. have.

A portion having a relatively low brightness on the converted image means a boundary with a relatively large change in chromaticity on the target image (A), and the step of calculating the anchor point of the target object (S40) is the target transformed image ( By converting C) to the resolution of the target image (A), the anchor point (P) of the target object (B) is calculated on a boundary with a large change in chromaticity, and displayed on the target object (B) have.

To explain this in more detail, as shown in Fig. 4, the step of calculating the anchor point of the target object (S40) is to convert the target transformed image (C) back to the resolution of the target image (A), Re-converting a portion with low brightness existing on the converted image (eg, a region having a boundary in which the contrast is greater than or equal to a preset standard on the converted target image C) into the resolution of the target image A At the same time as corresponding to the target object D, the target object in which the center point of a region (a plurality of blocks) having a boundary in which the contrast is different than a preset standard on the target transformation image C is re-transformed An anchor point (P) can be created on the re-transformed target object (D) by displaying the corresponding part on (D).

Here, as an example, the method for detecting a mosaic image may further include calculating vertical and horizontal edges of the target object based on the calculated anchor point P of the target object B ( S50 ).

To explain this in more detail, FIG. 5 shows that the vertical and horizontal edges (yellow lines) calculated based on the anchor point P and the anchor point (red point, blue point) are displayed on the target object B. As one thing, the step of calculating the vertical and horizontal edges of the target object (S50) is a boundary in which the contrast and/or chromaticity are different than a preset standard based on the anchor point P of the target object B. , and the calculated boundary is defined as the vertical and horizontal edges of the target object B as the vertical and horizontal edges of the target object B as the vertical and horizontal edges that are greater than or equal to the preset distance based on the anchor point P of the target object B. It can detect vertical and horizontal edges.

A detailed description of the method for detecting such vertical and horizontal edges will be omitted since various techniques using the characteristics of the mosaic described above are already known.

Meanwhile, the method for detecting a mosaic image may further include determining whether the target object is a mosaic by using the calculated horizontal and vertical edges of the target object B ( S60 ).

To explain this in more detail, the step of determining whether the target object is a mosaic ( S60 ) may calculate the number of vertical and horizontal edges shown in FIG. 5 , and the calculated number of vertical and horizontal edges is a preset number. In the above case, the target image A (the target object B) may be defined as a mosaic-processed image, and accordingly, the corresponding image may be defined as a mosaic-processed image.

Conversely, in the step S60 of determining whether the target object is a mosaic, if the calculated number of vertical and horizontal edges is less than a preset number, the target image A (the target object B) is not mosaic. It can be defined as a non-mosaic image, and therefore, the corresponding image can be defined as an image without mosaic processing.

Meanwhile, there are cases in which it is unclear whether the corresponding image is a mosaic-processed image due to the above-described steps.

This case may occur when the corresponding image is of very low quality.

Accordingly, the method for detecting a mosaic image may further include the following steps.

As an example, the mosaic image detection method includes selecting a comparison area other than the target object B on the target image A ( S70 ), and selecting the comparison area E rather than the resolution of the target image A generating a comparison transformed image by converting it to a low preset resolution (S80), calculating an anchor point P (anchor point) of the comparison region E on the comparison transformed image (S90) and the calculated The method may further include calculating vertical and horizontal edges of the comparison area E based on the anchor point P of the comparison area E ( S100 ).

To explain this in more detail, as shown in FIG. 6 , in the mosaic image detection method, an area other than the target object B selected by the step S20 of selecting the target object on the target image A After selecting the comparison area E, in the same way as the target object B described above with reference to FIGS. 4 and 5, generating the comparison conversion image for the comparison area E (S80) ), calculating the anchor point P of the comparison area E (S90) and calculating the vertical and horizontal edges of the comparison area E (S100) may be implemented.

That is, the step (S80) of generating the comparison transformed image corresponds to the step (S30) of generating the target transformed image, and the step (S90) of calculating the anchor point P of the comparison area (E) is the Corresponding to the step (S40) of calculating the anchor point of the target object, the step of calculating the vertical and horizontal edges of the comparison area (E) (S100) is a step of calculating the vertical and horizontal edges of the target object (S50) can be matched with

Here, the step (S60) of determining whether the target object is a mosaic comprises comparing the calculated horizontal and vertical edges of the comparison area (E) with the horizontal and vertical edges of the target object (B) to obtain the target image ( It can be determined whether A) is a mosaic-processed image.

That is, in the step S60 of determining whether the target object is a mosaic, as described above, the calculated number of horizontal and vertical edges on the target object B, furthermore, the calculated horizontal and vertical edges on the comparison area E When the number of horizontal and vertical edges on the target object B calculated by comparing the number of vertical edges with each other is greater than or equal to a preset value, the target image A ) can be determined as a mosaic-processed image, and conversely, when the calculated number of horizontal and vertical edges on the target object B is less than a preset value than the calculated number of horizontal and vertical edges on the comparison area E, the It may be determined that the target image A is an image in which the mosaic is not processed.

Accordingly, it can be determined whether the mosaic has been accurately processed even when the target image A itself is of low quality.

Meanwhile, in the selecting of the comparison area ( S70 ), the comparison area E having the same size as the area occupied by the target object B on the target image A may be selected.

To implement this, the step of selecting the comparison area (S70) includes a plurality of divided areas having the same size with respect to at least a portion of the area excluding the area occupied by the target object B on the target image A ( a red box) and selecting a plurality of the divided areas to select the comparison area E having the same size as the area occupied by the target object B.

To explain this in more detail, as shown in FIG. 6 , for example, when the size of the area occupied by the target object B is 128x128, the step of selecting the comparison area (S70) is the target object B. For an area excluding the area of , a plurality of divided areas having a size of 8x8 are generated, and 16 divided areas are selected to select the comparison area E equal to the size of the area occupied by the target object B /can be defined

Thereafter, the mosaic image detection method may determine whether the corresponding image is a mosaic-processed image by detecting horizontal and vertical edges through the method described above for the selected comparison region E. FIG.

In the above, the configuration and features of the present invention have been described based on the embodiments according to the present invention, but the present invention is not limited thereto, and it is understood that various changes or modifications can be made within the spirit and scope of the present invention. It is intended that such changes or modifications will be apparent to those skilled in the art, and therefore fall within the scope of the appended claims.

100: control unit
200: memory unit
300: communication department
400: input unit
500: output unit

Claims (8)

A mosaic image detection method for detecting a mosaic in an image using a mosaic detection system, the method comprising:
selecting a target image, which is at least one of a plurality of images constituting an image;
selecting a preset target object on the target image;
generating a converted target image by converting the target object to a preset resolution lower than the resolution of the target image;
calculating an anchor point of the target object on the target transformation image;
calculating vertical and horizontal edges of the target object based on the calculated anchor point of the target object; and
Determining whether the target object is a mosaic using the calculated horizontal and vertical edges of the target object;
Mosaic image detection method.
According to claim 1,
Calculating the anchor point of the target object comprises:
Converting the target conversion image to the resolution of the target image to calculate the anchor point of the target object,
Mosaic image detection method.
3. The method of claim 2,
Calculating the anchor point of the target object comprises:
defining an anchor point of the target object as a center point of a region having a boundary in which contrast is different than a preset reference on the target transformation image,
Mosaic image detection method.
4. The method of claim 3,
Calculating vertical and horizontal edges of the target object includes:
A boundary having a difference of greater than or equal to a preset standard is calculated based on the anchor point of the target object, and vertical and horizontal edges of the target object are defined when the boundary is greater than or equal to a preset distance based on the anchor point of the target object. doing,
Mosaic image detection method.
5. The method of claim 4,
Determining whether the target object is a mosaic comprises:
When the calculated vertical and horizontal edges of the target object are greater than or equal to a preset number, the target image is defined as a mosaic,
Mosaic image detection method.
6. The method of claim 5,
selecting a comparison area other than the target object on the target image;
generating a comparison conversion image by converting the comparison area to a preset resolution lower than the resolution of the target image;
calculating an anchor point of the comparison region on the comparison transformed image;
Calculating vertical and horizontal edges of the comparison region based on the calculated anchor point of the comparison region; further comprising,
Determining whether the target object is a mosaic comprises:
Comparing the calculated horizontal and vertical edges of the comparison area with horizontal and vertical edges of the target object to determine whether the target image is a mosaic,
Mosaic image detection method.
7. The method of claim 6,
The step of selecting the comparison area comprises:
selecting the comparison area having the same size as the area occupied by the target object on the target image;
Mosaic image detection method.
8. The method of claim 7,
The step of selecting the comparison area comprises:
A plurality of divided regions having the same size are generated for at least a portion of the region excluding the region occupied by the target object on the target image, and an area having the same size as the region occupied by the target object by selecting a plurality of the divided regions branch to select the comparison area,
Mosaic image detection method.

Images