KR20030093825A - Method and apparatus for embedding watermark in 3 dimensional image by using virtual distance image - Google Patents

Abstract

PURPOSE: A method and an apparatus for inserting a watermark of a three-dimensional image using a virtual distance image are provided to insert a watermark into a three-dimensional image by using a virtual distance image and a standard coordinate system regardless of rotation or translation of the three-dimensional image. CONSTITUTION: Information of a three-dimensional image into which a watermark is inserted is inputted to a standard coordinate system setting unit(10). The standard coordinate setting unit sets up a standard coordinate system from the information of the three-dimensional image by using an eigenvector of a covariance matrix of the three-dimensional image. A scan coordinate system is set up from the standard coordinate system to form a cylindrical screen set up in a circumference of the three-dimensional image. A virtual distance measuring unit(20) measures a distance from the screen to the three-dimensional image and unfolds the screen in plane for forming a virtual distance image. A watermark inserting unit(30) inserts the watermark into the virtual distance image. A three-dimensional image reconstructing unit(40) reconstructs the virtual distance image as a three-dimensional image.

Description

Method and apparatus for embedding watermark of 3D image using virtual distance image {METHOD AND APPARATUS FOR EMBEDDING WATERMARK IN 3 DIMENSIONAL IMAGE BY USING VIRTUAL DISTANCE IMAGE}

The present invention relates to a method and apparatus for embedding a watermark on a 3D image, and more particularly, to a water for a 3D image using a virtual distance image using a reference coordinate system that is not affected by rotation or translation of the 3D image. A mark insertion method and apparatus.

Recently, various forms of digital works have been created according to the provision and distribution of information using the Internet. However, digital works have many problems in copyright protection due to their characteristics. It is easy to store or convert information, so that a large amount of copying is possible in open spaces such as the Internet without any restrictions, and it is possible to maintain the same data information as the original even after copying and to distribute it without the author's consent, thus protecting the copyright of the author. Difficult to do

This problem of copyright infringement is increasing in seriousness, and many legal, institutional and technical solutions have been devised. To this end, encryption is performed on the data so that only specific users can use the information, and digital watermarking hides copyright information from the data and delivers it to the user so that the data is not externally modified. Watermarking) is a technical solution that can solve the above-mentioned problems.

Among them, watermarking is a copyright protection technology for digital data, and may be referred to as a copyright display and verification technology for digital content. By concealing copyright information into data and re-detecting hidden copyright information, it is possible to detect data tampering and forgery, and to claim the author's ownership.

In this watermark embedding method, a spatial method, which is a method of adding a small change to data such as a pixel value of a screen and using it as a watermark, converts digital data into an analog signal of frequency component, Frequency domain method using discrete cosine transform (DCT), fast fourier transform (FFT), or wavelet transform Etc. This method is divided into visible, ie, visible watermarking technology and invisible watermarking technology.

Visible watermarking technology is a form of overlaying copyright information on the original, which generally prevents the user from erasing the copyright information. However, it is difficult to preserve the original value because the copyright information is visible on the original. Technically, visible watermarking technology is easy to implement, but it is difficult to use for e-commerce because the content is damaged.

On the other hand, invisible watermarking technology is invisible, so it is possible to protect the copyright without damaging the original, and if a copyright problem occurs, it is possible to extract it and claim its copyright based on it. Currently, most watermarking techniques use invisible methods. Therefore, when invisible watermarking is used, even if images, audio, and video files are distributed over the Internet, users cannot detect whether or not a watermark is in the distributed work.

As for the embedding and extraction of watermarks for the digital contents as described above, the applicant has already proposed various forms of methods in Patent Application Nos. 2000-01096, 2001-01940, 2001-84207, etc. have.

However, the above-mentioned watermark insertion / extraction method is for 2D images, and it is not directly applicable to 3D images, which are widely used in CAD, animation, and virtual reality, which are widely used recently. . As a result, researches have been actively attempting to use information as a basis for copyright by inserting information that can be decoded only by the author in the 3D image.

A widely used method of expressing 3D images is a mesh structure. The mesh structure basically consists of vertices and information connecting them. In order to insert information into the mesh structure, some modifications are made to the vertices and the connection information. However, even in the same model, the three-dimensional image has an infinite combination of expression methods, that is, coordinates of vertices and connection information. It is also possible to convert rotations and translations. It is a big challenge to ensure that hidden information (watermarks) embedded in such a high degree of freedom 3D image survives the transformations and intentional attacks that are routinely applied.

In particular, when a watermark is inserted in a rectangular coordinate system based on a coordinate point with respect to the 3D image, watermarking the direction of the 3D image to extract the watermark when the 3D image is rotated or translated. A reorientation process is needed to bring it back to the state it was in. However, this reorientation process requires a lot of information about the 3D image. Furthermore, the reorientation process for the 3D image has a disadvantage in that the extraction of the watermark requires the existence of the original in which the watermark is inserted.

Accordingly, an object of the present invention is to provide a method and apparatus for embedding a watermark in a 3D image.

Another object of the present invention is to provide a method and apparatus for embedding a watermark in a 3D image using a virtual distance image.

Another object of the present invention is to provide a method and apparatus for inserting a watermark into a 3D image using a virtual distance image using a reference coordinate system irrelevant to transformation such as rotation or translation of the 3D image.

Another object of the present invention is to estimate the pose of the original using the eigenvector of the covariance matrix of the vertex coordinates for the three-dimensional image by using a virtual distance image that allows to insert / extract the watermark without the original A method and apparatus for embedding a watermark in an image are provided.

1 is a block diagram showing a schematic configuration of an apparatus for inserting a watermark of a 3D image using a virtual distance image according to the present invention.

2 is a view showing a state in which a cylindrical screen is installed around a three-dimensional image.

3 shows a relationship for measuring the distance between a screen and an arbitrary grid point in a scan coordinate system.

※ Explanation of codes for main parts of drawing

10 ... reference coordinate setting unit 20 ... virtual distance measuring unit

30 ... watermark inserter 40 ... 3D image reconstructor

In order to achieve the above object, a method for inserting a watermark into a 3D image using a virtual distance image according to the present invention is provided.

Setting a reference coordinate system using the eigenvectors of the covariance matrix of the 3D image from the input 3D image information;

Setting a scan coordinate system for forming a cylindrical screen set around the three-dimensional image from the reference coordinate system;

Calculating a distance from the plurality of grid points partitioning the screen at predetermined intervals to the three-dimensional image and spreading the screen in a plane to form a virtual distance image;

Inserting a watermark into the virtual distance image; And

And reconstructing the virtual distance image into which the watermark is inserted into a 3D image.

In addition, the apparatus for inserting a watermark in a three-dimensional image using a virtual distance image according to the present invention

Means for setting a reference coordinate system from the input information of the 3D image by using the eigenvector of the covariance matrix of the 3D image;

Means for setting a scan coordinate system for forming a cylindrical screen around the three-dimensional image from the set reference coordinate system;

Means for calculating a distance from the screen to the three-dimensional image and spreading the screen in a plane to form a virtual distance image;

Means for embedding a watermark in the virtual distance image; And

And means for reconstructing the virtual distance image into which the watermark is inserted into a 3D image.

Hereinafter, a method and system for inserting a watermark into a 3D image using a virtual distance image according to the present invention will be described in detail with reference to the accompanying drawings.

First, a schematic configuration of an apparatus for embedding a watermark into a 3D image using a virtual distance image according to the present invention will be described with reference to FIG. 1.

As shown in FIG. 1, the apparatus for inserting a watermark of a 3D image using a virtual distance image according to the present invention sets a reference coordinate system for inserting a watermark using coordinates of a vertex of a 3D image having a mesh structure. The reference coordinate system setting unit 10 for the virtual distance measuring unit 20 for measuring the distance from the specific point on the screen to the three-dimensional image after setting the cylindrical screen around the three-dimensional image on the set reference coordinates, Reconstruct the watermark inserting unit 30 for inserting the watermark into the two-dimensional virtual distance image generated by the virtual distance measuring unit 20 and the two-dimensional virtual distance image into which the watermark is inserted into a three-dimensional plane. 3D image reconstructing unit 40 is included.

The operation of each component in the watermark embedding apparatus of the 3D image having the above configuration will be described in detail.

First, the reference coordinate system setting unit 10 receives information of a 3D image into which a watermark is to be inserted. The information of the 3D image is composed of three-dimensional coordinates of the vertices and connection information indicating how the vertices are connected to each other. In the xyz coordinate system, which is a coordinate used by the information of the 3D image, it is difficult to restore the original coordinates when the 3D image is rotated or translated based on a specific coordinate in space. Therefore, even when a change in coordinates is caused by rotation or translation, it is necessary to set a coordinate that can always be a reference.

In order to set the reference coordinates, the reference coordinate system setting unit 10 sets the reference coordinate system using the coordinates of the vertices provided by the information of the 3D image. In this case, a covariance matrix of vertex coordinates is obtained and a eigenvector is calculated from the matrix (Principal Component Analysis; PCA).

The reference coordinate system setting unit 10 sets the reference coordinate system through the following process. First, we obtain the covariance matrix using the x-, y-, and z-axis coordinates of the vertices of the given 3D image and set the eigenvectors of the matrix as the three coordinate axes of the reference coordinate system. If the number of vertices of the 3D image is N, the covariance between two coordinate components is expressed by Equation 1.

Where x [i] and y [i] are the x and y coordinates of the i th vertex, and mx and my are the mean of the x and y axis components. By using Equation 1, a covariance matrix represented by Equation 2 with respect to the coordinates of a 3D image may be obtained.

The eigenvectors (e ₁ , e ₂ , e ₃ ) can be obtained from the covariance matrix obtained as described above. The eigenvectors thus obtained have the characteristic of being rotated to the same degree when the 3D image is rotated.

For example, if the rotation transformation matrix is R, when all the vertices v of the 3D image are as follows when the eigenvectors of the covariance matrix C of the 3D image are (e ₁ , e ₂ , e ₃ ) Rotating with v '= Rv, the eigenvector of the rotated 3D image is rotated to the same degree as the 3D image with e ₁ ' = Re ₁ , e ₂ '= Re ₂ , e ₃ ' = Re ₃ . Therefore, these eigenvectors are used as the basis for the scanning cylinder, which will be described later, and the cylinder can always be set at the same position no matter how the 3D image is rotated or translated. The concealed hidden information can always be extracted.

When the reference coordinate system represented by the eigenvectors (e ₁ , e ₂ , e ₃ ) is set, the key code for expressing the hidden information (watermark) to be inserted is a key code for creating a cylindrical screen. Set the three axes s ₁ , s ₂ , and s ₃ . In this description, the scan coordinate system is the same method as scanning a 3D image using a laser in a 3D scanner in the following description, and the coordinate system used at this time is referred to as a 'scan coordinate system'. The three axes of the scan coordinate system are three-dimensional vectors: s ₁ = (s ₁ a, s ₁ b, s ₁ c), s ₂ = (s ₂ a, s ₂ b, s ₂ c), s ₃ = (s ₃ a , ₃ s b, s c ₃₎ is represented by the three vectors of these it is perpendicular with respect to each other, and the size is 1, and determines the direction of the center axis and the radial direction of the scan coordinate system. In addition, the three axes of the scan coordinate system are vectors based on the e ₁ , e ₂ , and e ₃ coordinate systems, not the x, y, and z coordinate systems. That is, one axis S = (s ₁ , s ₂ , s ₃ ) of the scan coordinate system may be expressed by Equation 3. S ₁ , s ₂ , and s ₃ , which constitute the scan coordinate system, are the keys for inserting / extracting watermarks.

As shown in Equation 3, the scan coordinate system is set to the coordinate (scan coordinate system) of the image to be scanned by the reference coordinate system obtained from the 3D image rather than the x, y, z coordinate system. Hidden information inserted in the dimensional image can be extracted.

In other words, if the x, y, z coordinate axes are used as the reference coordinate system, when the hidden information is inserted by scanning based on the y axis and the object is rotated and the y axis is changed, the watermark is no longer extracted based on the y axis. S ₁ , s ₂ , and s ₃ , where s ₁ is the central axis of the cylinder and s ₂ and s ₃ are the radial directions of the cylinder, form the cylindrical screen that forms the scan coordinate system. Created based on eigenvectors.

As described above, the eigenvectors rotate with the rotation of the three-dimensional model. Thus, no matter how the 3D model rotates, it always has the same position with respect to the model. And since the center of the three-dimensional model (average of all vertex coordinates) is the center of the cylindrical screen coordinate system, when the three-dimensional model translates, the cylinder moves together so that the inserted watermark can be extracted regardless of the translation.

When the scan coordinate system is determined by the reference coordinate system obtained by the reference coordinate system setting unit 10, the virtual distance measuring unit 30 performs a virtual operating method of the 3D scanner to obtain a virtual distance image by measuring the virtual distance. The scan coordinate system generated by the reference coordinate system setting unit 10 using the key information and the eigenvector is used as the coordinate system of the virtual distance measuring unit 30. First, a screen of a cylinder is set around a 3D image based on the scan coordinate system as shown in FIG. 2. At this time, the radius (R) of the cylinder and the height (L) of the cylinder are set on the cylindrical screen so that the entire 3D image enters the cylinder.

Height (L) of the cylinder may obtain the center axis direction of the furthest vertex (s ₁₎ upwards in the direction of the most distant corner and bottom (s direction ₁₎ (-s ₁ direction) obtained by calculating the distance, a cylindrical The radius R of is obtained by obtaining the radial component of each vertex and setting the largest one as the radius.

After setting, the screen is divided into horizontal V and vertical U grids, and the distance from the intersection of each grid to the 3D image is obtained. It shoots a virtual light at each grid point to find the point where it meets the plane of the 3D image (triangle in the triangular mesh) and the distance from the grid point (u, v) to the intersection point.

Once the distances are found at all intersections, the screen can be unfolded to obtain a virtual distance image of size V x U. At this time, the position of each grid point becomes the coordinate value of the pixel in the image and the measured distance plays a role similar to the color of each pixel. This is because the two-dimensional image has 0 to 255 color values, but the virtual distance image has the measured distance value, and thus this value plays a role similar to the color of the pixel.

A method of measuring a distance at each grid point u and v will be described with reference to FIG. 3. In the scan coordinate system, the s1 axis is set as the center axis of the cylinder in the x, y, z coordinate system, and the origin of the scan coordinate system is (mx, my, mz) which is the average point of each vertex in the x, y, z coordinate system. The Eu vector is the central axial direction of the cylinder and its magnitude is the distance between the grids in the U direction. The Ev vector is a direction perpendicular to the central axis and pointing to the lattice point in the central axis, the magnitude of which is 1. Equation 4 below is an expression of Eu and Ev vectors, respectively.

When the straight line in the direction of Ev passing through the grid point (u, v) meets the plane of the 3D image, the distance between the point and the grid point (u, v) becomes the distance to be measured. In this way, the distance is calculated for all grid points. In the present invention, since a three-dimensional mesh structure is used, each surface is a polygonal surface formed by connecting vertices, and thus, the distance is calculated from the specific points of the cylindrical screen to the intersection of the polygonal surface of the three-dimensional model.

The watermark inserting unit 30 is a process of inserting concealment information into the colossal distance image. The virtual distance image is converted into a form of 2D image. That is, in the case of the conventional 2D image, the coordinates and the color values at the corresponding coordinates are expressed, but the virtual distance image according to the present invention is composed of coordinates based on grid points and distance values at the corresponding coordinates. Considering the color value, it can be regarded as the same form. Therefore, the two-dimensional image watermarking used in the watermark inserter 30 has been actively studied to date and can be achieved by applying various techniques.

As such, when a watermark is inserted into the 2D virtual distance image, the image is reconstructed into the original 3D image. The 3D image reconstructing unit 40 converts distance information at grid points u and v into coordinates of 3D vertices by performing a reverse process of virtual distance measurement. The vertices are then joined together three times to form a triangular surface.

That is, since the coordinates in the virtual distance image with the watermark correspond to the coordinates of the grid points, the coordinates are converted into the cylindrical screen using the coordinates. After the conversion, as shown in FIG. 3, the distance values in which the watermark is inserted from the grid point of the cylindrical screen in the center direction correspond to the vertices of the triangle, and these vertices are connected to each other to form a surface having a triangular face. If you make it, you can get 3D image with watermark embedded in the original 3D image.

As described above, the present invention may conceal and transmit information in a 3D image in which a 3D image is formed of a vertex and a connection thereof, that is, a 3D image formed of a polygonal vector image. In addition, the present invention does not use a coordinate system including coordinate information about a vertex of a 3D image, but sets a separate reference coordinate system that is not affected by deformations such as rotation or translation of the 3D image, and thus watermarks. There is an effect of facilitating extraction of the watermark by inserting.

While the invention has been particularly shown and described with reference to the above embodiments, it has been used for the purpose of illustration and those of ordinary skill in the art, having the spirit and scope of the invention as defined in the appended claims. Various modifications can be made without departing.

Claims (15)

Setting a reference coordinate system using the eigenvectors of the covariance matrix of the 3D image from the input 3D image information; Setting a scan coordinate system for forming a cylindrical screen set around the 3D image from the reference coordinate system; Calculating a distance from the screen to the 3D image and spreading the screen to a plane to form a virtual distance image; Inserting a watermark into the virtual distance image; And And reconstructing the virtual distance image into which a watermark has been inserted into a three-dimensional image. The method of claim 1, wherein the information of the 3D image includes 3D coordinates of vertices and connection information of the vertices. The method of claim 2, In the setting of the reference coordinate system, a covariance matrix is obtained from coordinates of vertices of the 3D image, and the eigenvector of the covariance matrix is set as a coordinate axis of the reference coordinate system, and the reference coordinate system is used to rotate or translate the 3D image. A watermark embedding method of a 3D image using a virtual distance image, which is not affected. The method of claim 3, wherein The origin of the scan coordinate system is a watermark embedding method of a three-dimensional image using a virtual distance image, characterized in that the average of the vertices of the three-dimensional image. The method of claim 4, wherein The scan coordinate system is a watermark insertion method of a three-dimensional image using a virtual distance image, characterized in that the center axis direction and the radial direction of the cylindrical screen. The method according to any one of claims 3 to 5, The virtual distance image forming step Forming a plurality of grid points formed by partitioning the screen horizontally and vertically at predetermined intervals; And Comprising a step of calculating the distance between the grid points and the 3D image watermark embedding method of a 3D image using a virtual distance image. The method of claim 6, In the cylindrical screen, the height of the cylinder is set by obtaining the farthest vertex upward and the farthest vertex downward in the direction of the central axis and calculating the distance, and the radius of the cylinder is obtained by obtaining the radial component of each vertex Watermark embedding method of a three-dimensional image using a virtual distance image, characterized in that set to. The method of claim 6, The 3D image reconstruction step Converting the virtual distance image into which the watermark is inserted into a cylindrical screen; And After the conversion, forming a surface having a triangular surface by connecting them to the vertices of the distance value at which the watermark is inserted in the center direction from the grid point of the cylindrical screen; A watermark embedding method of a three-dimensional image using a virtual distance image, comprising reconstructing a three-dimensional image having a watermark inserted into an original three-dimensional image. Means for setting a reference coordinate system based on the eigenvectors of the covariance matrix of the 3D image from the input 3D image information; Means for setting a scan coordinate system for forming a cylindrical screen around the three-dimensional image from the set reference coordinate system; Means for calculating a distance from the screen to the three-dimensional image and spreading the screen in a plane to form a virtual distance image; Means for embedding a watermark in the virtual distance image; And And a means for reconstructing the virtual distance image into which a watermark has been inserted into a three-dimensional image. 10. The apparatus of claim 9, wherein the information of the 3D image includes 3D coordinates of vertices and connection information of the vertices. The method of claim 10, The means for setting the reference coordinate system obtains a covariance matrix from the coordinates of the vertices of the 3D image and sets the eigenvector of the covariance matrix as the coordinate axis of the reference coordinate system, wherein the reference coordinate system corresponds to rotation or translation of the 3D image. Watermark embedding apparatus of a three-dimensional image using a virtual distance image, characterized in that it is not affected. The method of claim 11, The scan coordinate system is formed in the direction of the central axis and the radial direction of the cylindrical screen, the origin of the scan coordinate system is a watermark of the three-dimensional image using a virtual distance image, characterized in that the average of the vertices of the three-dimensional image Insertion device. The method according to claim 11 or 12, The virtual distance image forming means Forming a virtual distance image by forming a plurality of grid points formed by partitioning the screen horizontally and vertically at predetermined intervals to calculate the distance between the grid points and the 3D image. Watermark embedding apparatus of 3D image using. The method of claim 13, In the cylindrical screen, the height of the cylinder is set by obtaining the farthest vertex upward and the farthest vertex downward in the direction of the central axis and calculating the distance, and the radius of the cylinder is obtained by obtaining the radial component of each vertex Watermark embedding apparatus of a three-dimensional image using a virtual distance image, characterized in that set to. The method of claim 13, The three-dimensional image reconstructing means converts the virtual distance image into which the watermark is inserted into a cylindrical screen, and connects the distance value where the watermark is inserted from the grid point of the cylindrical screen to the center by a vertex. 3. A watermark insertion apparatus for a three-dimensional image using a virtual distance image, wherein the surface having a triangular surface is formed to reconstruct a three-dimensional image in which a watermark is inserted into the original three-dimensional image.