JP2007140820A - Creation method for texture data for rendering - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To simultaneously materialize both compression ratio improvement and quality maintenance when compressing texture data of a high-dimensional image. <P>SOLUTION: A function BTF (θL, ϕL, θV, ϕV, u, v) having six variables of coordinate values u, v in uv two-dimensional coordinate system, an incidence angle θL and an azimuth angle ϕL of illumination light, and an emission angle θV and an azimuth angle ϕV of reflection light is prepared as the texture data 20. In each pixel T (u, v), a table R (u, v) showing a reflection characteristic is defined in each combination of four angles θL, ϕL, θV, ϕV. A part of all the pixels is decided as a representative pixel, and the remainder is set as a general pixel. About the representative pixel, contents of the table R (u, v) are used as they are. About the general pixel, a row or a column related to ϕL or ϕV of the table R (u, v) is performed with a rotation shift, tables having a plurality of variations are derived in each the representative pixel, and the most approximate table is substituted. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a method for creating rendering texture data, and more particularly, to a compression technique for high-dimensional texture data in which reflection characteristics are determined depending on an angle with respect to an incident direction or a reflection direction of light.

  Due to the improvement in computer performance, CG images are used in various fields of industry. For example, in the design stage of buildings, furniture, automobiles, etc., many CG images are usually used. Also, various CG images of articles are indispensable for various video expressions such as product presentations and movies using computers. Furthermore, recently, CG images are often used instead of actual product photos in product catalogs and the like.

  Article data such as buildings, furniture, and automobile interior parts is three-dimensional data of a virtual object, whereas a CG image for presentation is usually prepared as a two-dimensional image. Therefore, when creating a CG image using 3D article data, the 3D data of the virtual object is taken into the computer, the illumination conditions, the viewpoint position, and the projection plane are set, and then the 3D virtual A rendering process for obtaining a two-dimensional projection image of the object on the projection plane is performed. This process basically simulates the phenomenon in which the illumination light from the light source determined by the illumination conditions is reflected at each part of the virtual object and travels to the viewpoint position on the computer. It can be said that the light intensity is calculated.

  Also, if you want to create an expression with a pattern or pattern pasted on the surface of a virtual object, you need to prepare two-dimensional texture data in advance and map this texture data to the surface of the object data during rendering. Done. Texture data is basically two-dimensional image data showing the distribution of reflection characteristics, and this is mapped to the surface of a virtual object to express an object having different characteristics in reflection characteristics for each part. It becomes possible.

  On the other hand, recently, a method using high-dimensional image data instead of simple two-dimensional image data as texture data has been proposed. For example, Non-Patent Document 1 below discloses a rendering process that uses texture data represented by a six-dimensional function called BTF (Bi-directional Texture Function). This BTF is a function that defines reflection characteristics that differ depending on the incident direction and reflection direction of light, and two angles θL and φL that indicate the incident direction of illumination light and two angles θV and φV that indicate the emission direction of reflected light. And a total of six variables (θL, φL, θV, φV, u, v), which are coordinate values u and v in the two-dimensional uv coordinate system, become a function that defines specific reflection characteristics. In order to take into account internal scattering near the surface of the material, BSSRDF (Bi-directional Surface Scattering Reflectance Distribution Function) has also been proposed as an eight-dimensional function to which variables x and y representing reflection points of multiple scattering are further added. ing. This BSSRDF becomes a function that defines a specific reflection characteristic by a total of eight variables (θL, φL, θV, φV, u, v, x, y).

Texture data given as such a high-dimensional image is usually prepared in the form of a numerical table, but the data capacity becomes enormous as the dimension increases. Therefore, a technique for compressing and using data for texture data composed of such a high-dimensional image has been proposed. For example, Patent Document 1 below discloses a technique for compressing the data amount by separating image information included in texture data into a light source-independent image and a luminance map. Patent Document 2 below discloses a technique for encoding and compressing texture data.
G. Muller, J. Meseth, M. Sattler, R. Sarlette and R. Klein 2004, Acquisition, Synthesis and Rendering of Bidirectional Texture Functions. In Eurographics 2004 STAR-State of The Art Report JP 2004-234285 A JP 2004-234476 A

  As described above, the texture data given as a high-dimensional image has an enormous amount of data. Therefore, it is essential to compress the texture data in order to handle it in a small computer system. However, in general, improvement of the compression ratio and quality maintenance are contradictory factors. If the compression ratio is increased, the quality of the texture data will be degraded, and if the quality is maintained, the compression ratio will be reduced. Become. In addition, when rendering is performed using compressed texture data, it is necessary to partially decompress the texture data. However, texture data obtained by a conventionally proposed compression method can be expanded at high speed. It is difficult to process.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a rendering texture data creation method that can perform high-speed expansion processing and can achieve both improvement in compression rate and quality maintenance as much as possible.

(1) The first aspect of the present invention is:
An original data preparation stage in which original texture data is prepared which includes a set of M pixels, and for each individual pixel, a table showing a reflection characteristic defined depending on an angle with respect to an incident direction and / or a reflection direction of light is defined. When,
A representative pixel determining step in which N (N <M) representative pixels are determined from the M pixels, and the remaining (MN) pixels are general pixels;
For each general pixel, an approximate representative pixel selection stage that selects a representative pixel having a reflection characteristic approximate to its own reflection characteristic as an approximate representative pixel;
Of the M pixels on the original texture data, N representative pixels are recorded with a table indicating their own reflection characteristics, and the (MN) general pixels are selected approximate representative pixels. A data compression stage for creating compressed texture data containing data indicating
In the method of creating rendering texture data having
At the approximate representative pixel selection stage, for each representative pixel, a table having a plurality of variations is defined by shifting the angle parameter of the table defined by the original texture data by a predetermined offset amount. The most approximate table when all the tables are compared is determined as the most approximate table, and the representative pixel from which this approximate table is derived is selected as the approximate representative pixel.
At the data compression stage, for general pixels, create compressed texture data that contains data indicating the selected approximate representative pixel and data indicating the offset amount for the most approximate table that caused the selection of the approximate representative pixel. It is what you do.

(2) According to a second aspect of the present invention, in the method for creating rendering texture data according to the first aspect described above,
In the original data preparation stage, a table is defined in which angles relating to the incident direction and / or reflection direction of light are expressed by a plurality of angle values obtained by equally dividing a predetermined range of angles,
At the approximate representative pixel selection stage, a shift is performed by adding or subtracting a predetermined offset amount corresponding to an integer multiple of the equally divided angle with respect to each angle value, and as a result of addition or subtraction, the angle value after the shift Rotate to count out the amount of protrusion from the other end of the predetermined range.

(3) According to a third aspect of the present invention, in the method for creating rendering texture data according to the second aspect described above,
At the original data preparation stage, on the surface of the virtual object to which the texture data is to be mapped, the sample point Q, the normal n set at the position of this sample point Q, and the sample point Q are orthogonal to the normal n. A reference plane S and a reference line ζ on the reference plane S passing through the sample point Q are defined, and an angle formed between the illumination light incident on the sample point Q and the normal n is θL, and illumination on the reference plane S The angle between the projected image of the light and the reference line ζ is φL, the angle between the reflected light emitted from the sample point Q and the normal n is θV, and the projected image of the reflected light on the reference plane S and the reference line ζ The original texture data is defined in which the reflection coefficient K (θL, φL, θV, φV) given as a function of “θL, φL, θV, φV” is defined as a table when the angle formed by φ is φV. Is.

(4) According to a fourth aspect of the present invention, in the method for creating rendering texture data according to the third aspect described above,
At the original data preparation stage, the angle range of 0 to 360 ° is equally divided by α to set the angle value of φL to α, the angle value of φV to α, and the angle range of 0 to 90 ° to β By equally dividing, the angle value of θL is set in β ways, the angle value of θV is set in β ways, and a table showing total α ² × β ² ways of reflection characteristics is prepared for each pixel. Is.

(5) According to a fifth aspect of the present invention, in the method for creating rendering texture data according to the fourth aspect described above,
In the approximate representative pixel selection stage, a unit angle Uφ given by the equation Uφ = 360 ° / α is defined, and φL and φV are defined as Δφ = i × Uφ (where i = 0, 1, 2,..., Α− 1) A table having a total of α ² combinations of variations each shifted by α offset amounts Δφ given by the following equation is defined.

(6) According to a sixth aspect of the present invention, in the method for creating texture data for rendering according to the fourth aspect described above,
In the approximate representative pixel selection stage, a unit angle Uφ given by an equation Uφ = 360 ° / α is defined, a unit angle Uθ given by an equation Uθ = 90 ° / β is defined, φL and φV are defined as Δφ = i × Uφ (where i = 0, 1, 2,..., α−1) are shifted by α different offset amounts Δφ given by the equation, and θL and θV are changed to Δθ = j × Uθ (where j = 0, 1, 2,..., β-1) so as to define a table having a total of α ² × β ² combinations of variations shifted by β offset amounts Δθ given by the following equation. It is a thing.

(7) According to a seventh aspect of the present invention, in the above-described rendering texture data creation method according to the first to sixth aspects,
In the original data preparation stage, a table having a numerical value indicating the reflectance for each of the illumination lights having the wavelengths of the three primary colors RGB is defined as data indicating the reflection characteristics.

(8) According to an eighth aspect of the present invention, in the method for creating texture data for rendering according to the seventh aspect described above,
In the approximate representative pixel selection stage, for the table showing its own reflection characteristics and the table to be compared, the table that minimizes the sum of squares of the corresponding individual reflectance differences is determined as the most approximate table. It is.

(9) According to a ninth aspect of the present invention, in the above-described rendering texture data creation method according to the first to eighth aspects,
At the original data preparation stage, an actual object is irradiated with illumination light from various directions, and images are taken from various directions, and original texture data is prepared based on a plurality of obtained images. It is a thing.

  (10) A tenth aspect of the present invention causes a computer to execute a representative pixel determination step, an approximate representative pixel selection step, and a data compression step in the rendering texture data creation method according to the first to eighth aspects described above. A program for this purpose is prepared.

  (11) In an eleventh aspect of the present invention, compressed texture data is created by the rendering texture data creating method according to the first to ninth aspects described above, and can be distributed.

(12) According to a twelfth aspect of the present invention, the compressed texture data created by the above-described rendering texture data creation method according to the first to ninth aspects is mapped onto the surface of a three-dimensional virtual object. In a rendering method for generating two-dimensional data indicating a projection image obtained when observing the image from a predetermined viewpoint,
An object data input stage in which a computer inputs object data indicating a structure of a three-dimensional virtual object;
A texture data input stage in which a computer inputs compressed texture data;
A rendering condition setting stage in which a computer sets a rendering condition for determining a lighting environment, a viewpoint position, a mapping mode, and a projection plane necessary for rendering based on an instruction from an operator;
A corresponding pixel determining step in which a computer determines, for each sample point Q on the three-dimensional virtual object, a corresponding pixel in the compressed texture data mapped to the position of the sample point Q;
When the corresponding pixel is a representative pixel, the computer calculates reflected light from the sample point Q using the reflection characteristics indicated by the table recorded for the representative pixel, and the corresponding pixel is a general pixel. In the case of, the table recorded for “approximate representative pixel indicated by the data recorded for the general pixel” is used as the “offset amount indicated by the data recorded for the general pixel”. The reflected light calculation stage for calculating the reflected light from the sample point Q using the reflection characteristics shown by the table obtained by shifting only by;
A two-dimensional data generation stage in which a computer generates two-dimensional data indicating a projection image obtained on a projection plane based on reflected light from each sample point Q;
Is to do.

(13) The thirteenth aspect of the present invention is
A set of M pixels. For each individual pixel, original texture data in which a table showing reflection characteristics determined depending on the angle of incidence and / or reflection direction of light is defined and stored. An original data storage unit,
A representative pixel determining unit that determines N (N <M) representative pixels from the M pixels and uses the remaining (MN) pixels as general pixels;
For each general pixel, an approximate representative pixel selection unit that selects a representative pixel having a reflection characteristic approximate to its own reflection characteristic as an approximate representative pixel,
Among M pixels, N representative pixels are recorded with a table indicating their own reflection characteristics, and (MN) general pixels are recorded with data indicating selected approximate representative pixels. A data compression unit for creating compressed texture data,
In the rendering texture data creating apparatus comprising:
The approximate representative pixel selection unit defines a table having a plurality of variations by shifting the angle parameter of the table defined by the original texture data by a predetermined offset amount for each representative pixel. The most approximate table when all the tables are compared is determined as the most approximate table, and the representative pixel from which this approximate table is derived is selected as the approximate representative pixel.
For general pixels, the data compression unit creates compressed texture data that contains data indicating the selected approximate representative pixel and data indicating the offset amount for the most approximate table that caused the selection of the approximate representative pixel. It is what you do.

(14) According to a fourteenth aspect of the present invention, the compressed texture data created by the rendering texture data creation device according to the thirteenth aspect described above is mapped onto the surface of a three-dimensional virtual object, In a rendering device that generates two-dimensional data indicating a projection image obtained when observed from a viewpoint,
An object data storage unit for inputting and storing object data indicating the structure of a three-dimensional virtual object;
A texture data storage unit for inputting compressed texture data and storing it;
A rendering condition setting unit for setting a rendering condition for determining a lighting environment, a viewpoint position, a mapping mode, and a projection plane necessary for rendering;
A corresponding pixel determining unit for determining a corresponding pixel in the compressed texture data mapped to the position of the sample point Q on the three-dimensional virtual object;
When the corresponding pixel is a representative pixel, the reflected light from the sample point Q is calculated using the reflection characteristics indicated by the table recorded for the representative pixel, and the corresponding pixel is a general pixel. In this case, the table recorded for the “approximate representative pixel indicated by the data recorded for the general pixel” is shifted by “the offset amount indicated by the data recorded for the general pixel”. A reflected light calculation unit that calculates the reflected light from the sample point Q using the reflection characteristics shown by the table obtained by
A two-dimensional data generation unit that generates two-dimensional data indicating a projection image obtained on the projection plane based on the reflected light from each sample point Q;
Is provided.

  According to the rendering texture data creation method of the present invention, it is possible to obtain compressed texture data that achieves both higher compression ratio and higher quality than ever before, and high-speed expansion processing is possible. Compressed texture data can be obtained.

Hereinafter, the present invention will be described based on the illustrated embodiments.
<<< §1. Basic concept of rendering processing >>
The present invention relates to a rendering process using a computer, and is particularly useful when applied to a rendering process for a three-dimensional virtual object in which a fiber sheet (a sheet made of a fibrous material such as cloth) is attached to the surface. It is. Here, for convenience of explanation, the basic concept of conventional general rendering processing will be described.

  FIG. 1 is a perspective view showing the principle of general rendering processing. This example shows a shading model for obtaining the two-dimensional projection image 15 on the projection plane H when the three-dimensional virtual object 10 is observed from the viewpoint E. At this time, the calculation for obtaining the projection image 15 is performed in a state where the texture data 20 prepared as a two-dimensional image is mapped to the surface of the virtual object 10.

  In order to perform such rendering processing, first, object data indicating the virtual object 10 is prepared. For example, in the illustrated example, the virtual object 10 is expressed as a set of a large number of polygons (polygons) defined on an XYZ three-dimensional coordinate system, and each part of the surface of the virtual object 10 is configured by a polygon. Has been. Therefore, the object data indicating the virtual object 10 is three-dimensional shape data composed of a collection of polygons, and the substance of the virtual object 10 is object data. Hereinafter, this will be referred to as virtual object 10 or object data 10 depending on the context of the description.

  On the other hand, the texture data 20 is two-dimensional image data defined on the uv two-dimensional coordinate system, and is an aggregate of a large number of pixels arranged on a two-dimensional plane. In the example shown here, a heart-shaped pattern is expressed. Of course, the pattern expressed by the texture data 20 is not necessarily an artificially drawn pattern, but may be an inherent shadow pattern of a natural material. For example, in the texture data expressing a plain carpet, an artificial pattern is not drawn, but a shadow pattern due to the fur of the carpet is expressed. After all, the texture data 20 can be referred to as reflection characteristic data indicating a distribution of reflection characteristics on a two-dimensional plane. When performing the rendering process, the texture data 20 defined on the uv coordinate system is mapped to the surface of the virtual object 10 defined on the XYZ coordinate system, and the reflection characteristics of each part of the virtual object surface are determined. It will be.

  The texture data 20 can be prepared, for example, by photographing a real fiber sheet using a scanner device, a digital camera, or the like, or artificially created on a computer using CG technology. You can also.

  In the rendering process, it is necessary to set the illumination conditions, the viewpoint position, and the projection plane. As illumination conditions, the type (for example, shape, size, wavelength characteristics, etc.) and position of the light source are set. In the case of the illustrated example, the light source G, the viewpoint E, and the projection plane H are set at the illustrated positions. Also, it is necessary to set a mapping mode (for example, relative position and orientation) indicating how the two-dimensional texture data 20 is mapped on the surface of the three-dimensional virtual object 10. On the other hand, an xy two-dimensional coordinate system is defined on the projection plane H, and the projection image 15 (a two-dimensional image obtained as an object of rendering processing) is obtained as image data on the xy coordinate system. become.

  Now, as shown in the figure, a sample point Q is defined on the virtual object 10, and the reflected light V from the sample point Q toward the viewpoint E is considered. The reflected light V is light obtained by reflecting the illumination light L from the light source G at the sample point Q, and is light that is observed at the position of the viewpoint E. However, “reflection” here is a broad concept that includes not only “specular reflection” but also “diffuse reflection”, and illumination light incident on the sample point Q at various angles is reflected as viewpoint E. Will head to. Accordingly, the reflected light V shown in the figure is a collection of reflected light obtained from illumination light incident from various angles.

  On the surface of the virtual object 10, a pixel P having a pixel value corresponding to the intensity of the reflected light V is defined at the intersection of the reflected light V from the one sample point Q toward the viewpoint E and the projection plane H. Similarly, if pixels are defined on the projection plane H with respect to each reflected light directed from the large number of sample points to the viewpoint E, the set of the large number of pixels defined for the large number of sample points results in the three-dimensional virtual object 10. A two-dimensional projection image 15 is obtained on the projection plane H. In practice, a pixel array is defined in advance on the projection plane H with a desired resolution, and a line connecting the viewpoint E and the representative point (for example, the center position) of each pixel P is extended onto the virtual object 10. If the sample point Q is defined at the intersection position with the virtual object surface and the pixel value of each pixel P is calculated by the above-described method, two-dimensional data having the necessary resolution can be obtained.

  As described above, in order to obtain the pixel value of the pixel P on the projection plane H, it is necessary to obtain the intensity of the reflected light V from the sample point Q through the pixel P to the viewpoint E. The intensity of V can be basically obtained by calculation in consideration of the intensity of the illumination light L incident on the sample point Q and the light reflectance (specular reflectance or diffuse reflectance) at the sample point Q. Of course, at this time, calculation taking into account the orientation of the polygon including the sample point Q is performed, and the texture data 20 (two-dimensional image data defined on the uv coordinate system) mapped to the sample point Q is also taken into consideration. The entered operation is performed. Therefore, if the surface of the virtual object 10 is a curved surface, a projection image 15 having a shadow corresponding to the curved surface is obtained, and the texture data 20 includes information on a heart-shaped pattern as shown in the figure. For example, this heart-shaped picture is also expressed on the projected image 15 (the picture is not shown in FIG. 1).

FIG. 2 is a perspective view showing the simplest method for calculating the intensity of the reflected light from the sample point Q in the rendering process shown in FIG. As shown in the figure, the intensity I (V) of the reflected light V from the sample point Q defined as one point on the XYZ three-dimensional coordinate system toward the viewpoint E is
I (V) = K · I (L)
It is expressed by the following formula. Here, I (L) is the intensity of the illumination light L at the position of the sample point Q, and K is the reflection coefficient. The intensity I (L) of the illumination light L is determined by parameters such as the brightness of the light source G, the distance between the light source G and the sample point Q, and the incident angle of the illumination light L with respect to the polygon to which the sample point Q belongs. On the other hand, the reflection coefficient K is determined by the texture data 20. Therefore, first, when the texture data 20 is mapped on the surface of the virtual object 10 in a predetermined mapping manner, the coordinate of the sample point Q corresponding to the coordinate of the texture data 20 is obtained and located at the corresponding coordinate. Pixel T (u, v) is determined. In FIG. 2, the position of the pixel T (u, v) corresponding to the sample point Q is indicated by an X mark. The value defined as the pixel value of the corresponding pixel T (u, v) becomes the reflection coefficient K at the position of the sample point Q.

  When a color image is used as the texture data 20, a predetermined reflection coefficient (reflectance) is defined for each primary color. FIG. 3 is a plan view showing a pixel array constituting the texture data 20. Here, for the sake of illustration, a very small pixel array of 6 rows and 8 columns is shown, but in practice, texture data having a larger pixel array is used. Each square in the figure represents one pixel (generally called “texel” in the sense of Texture Cell), and the pixel T (u, v) is represented by a coordinate value (u, v). Corresponds to the position pixel. In this example, three coefficients Kr, Kg, and Kb are defined as pixel values of one pixel T (u, v). Here, the coefficient Kr is the reflectance for the illumination light having the wavelength component of the primary color R (red), the coefficient Kg is the reflectance for the illumination light having the wavelength component of the primary color G (green), and the coefficient Kb is the primary color B. It is a reflectance with respect to illumination light having a wavelength component of (blue).

When mapping the texture data 20 composed of such a color image, the intensity of the reflected light from the sample point Q is calculated for each wavelength component of the light. For example, the intensity Ir (V) of the red wavelength component of the reflected light V is calculated using the intensity Ir (L) at the position of the sample point Q of the red wavelength component of the illumination light L and the reflection coefficient Kr.
Ir (V) = Kr · Ir (L)
It can be calculated by the following formula.

<<< §2. Model using high-dimensional image as texture >>
In §1, an example using a two-dimensional image as the texture data 20 is shown. In such a model for mapping a two-dimensional image, the reflection characteristic of the sample point Q is handled as not depending on the incident angle of illumination light or the emission angle of reflected light, but a three-dimensional structure such as a pile ground is handled. Strictly speaking, when the fiber sheet is stuck to the object surface, it must be handled in consideration of the anisotropic reflection characteristic depending on the incident angle of the illumination light and the emission angle of the reflected light.

  As described above, a model called BRDF (Bi-directional Reflectance Distribution Function) has been proposed as a basic model for handling with different reflection characteristics depending on the angle of incidence and reflection direction of light. . FIG. 4 is a perspective view for explaining the principle of the BRDF model. As shown in the figure, a normal line n is set at the position of the sample point Q on the virtual object surface, and an orthogonal plane S (hereinafter referred to as a reference plane S) including the sample point Q and orthogonal to the normal line n is defined. An angle formed between the normal line n and the incident illumination light L is θL, a projection image L ′ of the incident illumination light L on the reference plane S, and a predetermined reference line ζ passing through the sample point Q on the reference plane S and Is defined as a combination of an angle θL and an angle φL. Similarly, the angle between the normal line n and the reflected light V is θV, the angle between the projected image V ′ of the reflected light V on the reference surface S and the reference line ζ is φV, and the direction of the reflected light V is the angle. It is defined by a combination of θV and angle φV. In other words, the angles θL and θV correspond to the elevation angles of the incident illumination light and the reflected light (the original “elevation angle” means an angle with respect to the horizontal plane, so strictly speaking, the angles θL and θV are “90 ° −elevation angle”. However, here, for the sake of convenience, this is simply referred to as “elevation angle”). The angles φL and φV correspond to the azimuth angles of incident illumination light and reflected light.

Here, when the intensity of the incident illumination light L at the sample point Q is I (L), the intensity I (V) of the reflected light V from the sample point Q is given as a function of θL, φL, θV, φV. Using the reflection coefficient K (θL, φL, θV, φV),
I (V) = K (θL, φL, θV, φV) · I (L)
The BRDF model is obtained by the following formula. The BRDF model is characterized in that the reflection coefficient K is given as a function of parameters “θL, φL” indicating the direction of the incident illumination light L and parameters “θV, φV” indicating the direction of the reflected light V.

A BTF (Bi-directional Texture Function) model described in the above-mentioned Non-Patent Document 1 or the like is obtained by adding the concept of texture mapping to the BRDF model. In the BTF model, coordinate values (u, v) on the two-dimensional uv coordinate system are further added as parameters to the four angle parameters used in the BRDF model, and the incident illumination light L at the sample point Q When the intensity is I (L), the intensity I (V) of the reflected light V from the sample point Q is
I (V) = K (θL, φL, θV, φV, u, v) · I (L)
It is given by That is, the intensity I (V) of the reflected light V shown in FIG. 4 relates to the angle from which direction the incident illumination light L is applied to the sample point Q and from which direction the reflected light V is emitted from the sample point Q. This is a factor that is determined depending on the parameter and also determined depending on the position (u, v coordinate value) of the sample point Q.

  FIG. 5 is a plan view showing the structure of the 6-dimensional texture data 20 used in this BTF model. As described above, in this BTF model, the reflection coefficient K is a parameter defined as a six-dimensional function K (θL, φL, θV, φV, u, v). This means 6-dimensional image data composed of a set of a large number of pixels arranged in a 6-dimensional space, and each pixel has a pixel value corresponding to a specific reflection coefficient K. However, for convenience of explanation, the 6-dimensional texture data 20 is illustrated as a two-dimensional pixel array on the uv plane as in the example shown in FIG. In this case, each pixel T (u, v) on this two-dimensional pixel array has a reflection coefficient K (θL, φL, θV, φV) determined as a function of four parameters “θL, φL, θV, φV”. Is defined respectively.

  Here, the reflection coefficient K (θL, φL, θV, φV) can be defined as some function expression including θL, φL, θV, and φV as variables, but in practice, it is defined as a numerical table. In general, the present invention is premised on the handling defined as a numerical table. In this way, when the reflection coefficient K is defined as a numerical table, the angle parameters θL, φL, θV, and φV cannot be handled as continuous quantities, and thus must be handled discretely.

  FIG. 6 shows an example of a discrete definition of the angle θL. As described above, the angle θL indicates an angle with respect to the normal line of the illumination light L incident on the sample point Q. In FIG. 6, illumination lights L1 to L6 that are incident on the sample point Q at six elevation angles are depicted. The incident angles θL1 to θL6 of these illumination lights L1 to L6 are 0 °, 15 °, 30 °, 45 °, 60 °, and 75 ° as illustrated. With respect to the incident angle (elevation angle), if six angles are discretely defined at intervals of 15 ° in this way, texture expression with a certain degree of accuracy can be achieved in practice.

  On the other hand, FIG. 7 shows an example of a discrete definition of the angle φL. As described above, the angle φL is an angle formed between the projection image L ′ of the illumination light L incident on the sample point Q onto the reference surface S and the reference line ζ (in this example, the direction of φL1 = 0 °). The azimuth angle of the illumination light L is shown. In FIG. 7, the azimuth angle of the illumination light incident on the sample point Q from 12 different directions is depicted. 7 corresponds to the reference plane S, and azimuth angles in the range of 0 ° to 360 ° are defined on the reference plane S every 30 °. With regard to the azimuth angle, if 12 angles are defined discretely at intervals of 30 ° in this way, texture expression with a certain degree of accuracy can be achieved in practice.

  As described above, the discrete definition examples of the angles θL and φL related to the incident illumination light L have been shown, but the same definitions can be made for the angles θV and φV related to the reflected light V. That is, regarding the angle θV, six discrete definitions of 0 °, 15 °, 30 °, 45 °, 60 °, and 75 ° are made, and the angle φV is an azimuth angle in the range of 0 ° to 360 °. May be defined every 30 °.

  If such a discrete definition is performed, the angles θL and θV each have six values, and the angles φL and φV each have twelve values. Therefore, the combination of these four angles is 6 That is, × 12 × 6 × 12 = 5184 ways. Accordingly, the reflection coefficient K defined for one pixel T (u, v) of the texture data 20 shown in FIG. 5 takes 5184 values depending on the combination of the respective angle values.

  These 5184 reflection coefficient K values can be prepared as a table R (u, v) indicating reflection characteristics. FIG. 8 is a plan view showing a configuration example of the texture data 20 defined using the table R (u, v) indicating such reflection characteristics. In this example, the texture data 20 is defined as a two-dimensional pixel array on the uv coordinate system, but each pixel T (u, v) has a table R (u, v) is defined. The table R (u, v) includes six angles θL (θL1 to θL6), twelve angles φL (φL1 to φL12), six angles θV (θV1 to θV6), and twelve angles φV (φV1). Three reflection coefficients Kr, Kg, and Kb are defined for all the combinations of (˜φV12) (5184 combinations). Here, each coefficient Kr, Kg, Kb is a value indicating the reflectance of each primary color wavelength component of the illumination light as described above.

  The rendering process using the BTF model using the texture data 20 as shown in FIG. 8 is performed as follows. First, as shown in FIG. 1, the texture data 20 is mapped on the surface of the three-dimensional virtual object 10 to determine a pixel T (u, v) corresponding to the position of the sample point Q. Next, by taking into account the position of the light source G and the orientation of the polygon to which the sample point Q belongs, the angles θL and φL of the illumination light L incident on the sample point Q, and the angle of the reflected light V emitted from the sample point Q θV and φV are determined. The direction of the reference line ζ may be determined in advance, for example, “u-axis mapping direction”. Then, with reference to the table R (u, v) showing the reflection characteristics for the pixel T (u, v) shown in FIG. 8, three reflection coefficients corresponding to combinations of specific angles θL, φL, θV, and φV. Kr, Kg, and Kb are obtained. Finally, the intensity value for each primary color component of the reflected light V is obtained based on these three reflection coefficients Kr, Kg, and Kb, and the pixel value of the pixel P on the projection plane H may be determined.

<<< §3. Conventional data compression method using vector quantization >>
In the case of a model that uses a high-dimensional image as a texture, such as the BTF model described in §2, the data capacity of the texture data 20 is enormous. The texture data 20 shown in FIG. 8 is a small-scale one having a pixel arrangement of 6 rows and 8 columns on the uv coordinate system for convenience of illustration, but a larger pixel arrangement is practically used.

  For example, consider the data capacity of texture data having a pixel array of 512 rows and 512 columns on the uv coordinate system. In this case, the total number of pixels defined on the two-dimensional pixel array is (512 × 512). In the case of a model using a two-dimensional image as a texture, it is only necessary to define the values of three reflection coefficients Kr, Kg, and Kb for each pixel. Therefore, if the value of one coefficient is expressed by 1 byte, The total data capacity of the texture data 20 is sufficient to be (512 × 512 × 3) bytes. This is a capacity that can be sufficiently expanded in the mounting memory space of a general personal computer.

  However, in a model using a high-dimensional image as a texture, such as the BTF model described above, it is necessary to accommodate, for example, a table R (u, v) showing reflection characteristics as shown in FIG. . Here, for the angles θL and θV, six discrete angles are defined as shown in FIG. 6, and for the angles φL and φV, twelve discrete angles are defined as shown in FIG. Assuming that the reflection coefficients Kr, Kg, and Kb expressed by 1 byte are defined for each of the four combinations of angles, the total data capacity of the texture data 20 is (512 × 512 × 6 × 12 × 6 × 12). X3) Bytes. This is a capacity exceeding 4 Gbytes, and in the case of a general personal computer, it is a capacity that is difficult to expand in the mounting memory space. Of course, if the size of the two-dimensional pixel array defined on the uv coordinate system is made larger or the discrete angle definition is made finer, the total data capacity of the texture data 20 further increases.

  Therefore, when the rendering process as shown in FIG. 1 is executed by a general system such as a personal computer, the texture data 20 composed of a high-dimensional image such as a BTF model cannot be used as it is. For this reason, in the above-mentioned Patent Documents 1 and 2 and the like, techniques for compressing and using the texture data 20 composed of high-dimensional images are proposed. For compression of high-dimensional images, a technique called vector quantization is usually used. For example, in the case of the BTF model shown in FIG. 8, the table R (u, v) showing the reflection characteristics for one pixel shown in the lower half of the figure is composed of four variables (θL, φL, θV, φV). Since all the combinations (5184 combinations) are tables that define specific reflection coefficients, respectively, mathematically, 1 is defined on a 5184-dimensional coordinate system having 5184 mutually orthogonal coordinate axes. It is equivalent to a book vector. Therefore, the texture data 20 having 48 pixels shown in the upper half of the figure can be expressed as 48 vectors defined on the 5184-dimensional coordinate system. Therefore, some of the 48 vectors are defined as representative vectors, and the other vectors are replaced with the most approximate representative vectors (representative vectors having the smallest Euclidean distance in the 5184-dimensional coordinate space). The total amount of data can be reduced. This is the basic concept of data compression by vector quantization.

  FIG. 9 is a plan view showing the compression principle of high-dimensional texture data based on the concept of such vector quantization. As shown in FIG. 8, the original texture data 20 before compression includes a table R (u, v) indicating reflection characteristics for all 48 pixels T (u, v) in total. On the other hand, in the compressed texture data 25 shown in FIG. 9, among the total of 48 pixels, four pixels T1, T2, T3, and T4 are defined as representative pixels, and the remaining 44 pixels are defined as general pixels. It has established. For the four representative pixels T1, T2, T3, and T4, tables R1, R2, R3, and R4 indicating reflection characteristics (tables equivalent to the table R (u, v) shown in FIG. 8) are recorded. However, for the remaining general pixels, for example, the i-th general pixel Ti, instead of recording the table Ri, approximate representative pixels (representing the representative pixels R1 to T4) including the table closest to the table Ri are recorded. Data indicating either) is recorded.

  For example, as a result of comparing the table Ri to be originally recorded in the general pixel Ti and the tables R1 to R4 recorded in the four representative pixels, respectively, when it is determined that the table R3 is closest, Data indicating the approximate representative pixel T3 is recorded in the general pixel Ti. If such a method is used, it is possible to create the compressed texture data 25 as shown in FIG. 9 based on the original texture data 20 as shown in FIG. 8, and the data capacity after compression is greatly increased. Can be reduced. That is, in the compressed texture data 25, it is sufficient to accommodate a table with a large data capacity for only the four representative pixels T1 to T4, and for the remaining 44 general pixels, only data indicating approximate representative pixels can be accommodated. Good.

  Since such compressed texture data 25 can be expanded in the mounting memory space of a general personal computer, rendering processing can also be practically performed on a personal computer. When it is necessary to refer to the position of the representative pixel in order to obtain the reflection characteristic at the position of the predetermined sample point Q using the compressed texture data 25, the reflection characteristic recorded in the representative pixel is indicated. Use the table as it is. On the other hand, when it is necessary to refer to the position of the general pixel, a table indicating the reflection characteristic recorded in the approximate representative pixel is recognized from the data recorded in the general pixel. Can be substituted.

  Of course, when such data compression is performed, the quality of the texture data itself is degraded. This is because, for the general pixel Ti, an approximate representative pixel table (for example, R3) is used instead of the original table Ri, and therefore approximate characteristics that are not the correct reflection characteristics are used. It is. In order to maintain the quality of the texture data, it is necessary to increase the ratio of the representative pixel to the total number of pixels. This is because as the number of representative pixels increases, the possibility of selecting an approximate representative pixel having a more approximate table with respect to the original table of general pixels increases. However, as the ratio of representative pixels increases, the compression rate decreases. Eventually, with the conventionally proposed compression methods, it is impossible to achieve both improvement in compression rate and quality maintenance. The present invention proposes a novel technique for solving such problems.

<<< §4. Basic focus of the present invention >>>
In the present invention, as typified by the texture data for the BTF model described in §2, a table indicating reflection characteristics determined depending on an angle such as a light incident direction and a reflection direction is defined for each pixel. An efficient compression method for texture data is proposed. First, the focus of this compression method will be described.

  Consider an actual fiber sheet 30 having a surface structure as shown in the side view of FIG. As shown in the figure, the row of fibers 31, 32, and 33 are implanted on the surface of the fiber sheet 30. For example, carpets and carpets are representative of the fiber sheet 30 as shown. Here, for convenience of explanation, the hair row fibers 31 planted in the region A on the fiber sheet 30 are inclined obliquely to the left, and the hair row fibers 32 planted in the region B portion. Is directed vertically upward, and the hair row fibers 33 planted in the region C are inclined obliquely to the right.

  Now, let us consider preparing the original texture data 20 as shown in FIG. 8 based on the actual fiber sheet 30 and compressing it to create the compressed texture data 25 as shown in FIG. .

  First, in order to create the original texture data 20, a large number of captured images are collected by irradiating the actual fiber sheet 30 with illumination light from various directions and photographing it from various directions. And based on the pixel value of each picked-up image regarding the same position on the fiber sheet 30, what is necessary is just to obtain | require the "reflection property depending on an angle" of the said position. For example, the reflection characteristic for the position indicated as T1 in FIG. 10 can be determined by the pixel value at the position corresponding to T1 in a large number of captured images collected under various imaging conditions. Thus, the tables R1, R2, Ri indicating the unique reflection characteristics are defined at the positions indicated by T1, T2, Ti in FIG. 10, respectively. The original texture data 20 is composed of a set of a large number of defined pixels.

  Subsequently, N (N <M) representative pixels are determined from the M pixels constituting the original texture data 20, and the remaining (MN) pixels are set as general pixels. Creates compressed texture data 25 by selecting approximate representative pixels. Here, it is assumed that the pixel corresponding to the position indicated by T1 on the fiber sheet 30 illustrated in FIG. 10 is selected as the representative pixel T1, and the pixel corresponding to the position indicated by T2 is selected as the representative pixel T2. , Let the pixel corresponding to the position marked Ti remain as a general pixel Ti. In this case, the tables R1 and R2 defined in the original texture data 20 are used as they are for the representative pixels T1 and T2, but the table Ri defined in the original texture data 20 is used for the general pixel Ti. Instead, the more approximate one of the tables R1, R2 for the representative pixels T1, T2 is substituted.

  In the example shown in FIG. 10, as described above, the hair row fibers 31 in the region A extend obliquely leftward, the hair row fibers 32 in the region B extend vertically upward, and the hair row fibers 33 in the region C extend obliquely rightward. . In such a case, it is generally considered that the table R2 for the representative pixel T2 has a higher degree of approximation to the table Ri for the general pixel Ti than the table R1 for the representative pixel T1. This is because the hair row fibers 33 in the region C extend obliquely in the right direction, whereas the hair row fibers 31 in the region A extend in the diagonal left direction, which is the opposite direction, and therefore the hair row fibers 32 that extend vertically upward. This is because it is considered that the degree of approximation regarding the reflection characteristics is higher in the region B having.

  FIG. 11 is a plan view showing the correlation between the tables R1, R2, Ri indicating the reflection characteristics defined in the original texture data 20 for the pixels T1, T2, Ti in these three regions A, B, C. is there. 8 illustrates the table R (u, v) for combinations of the four angles θL, φL, θV, and φV, but in FIG. 11, the elevation angles θL and θV are set to specific angles for convenience of illustration. As fixed, only the table for the combination of azimuth angles φL and φV will be shown. In other words, the tables R1, R2, Ri shown in FIG. 11 are partial tables in which only the portions corresponding to θL1, θV1 in the table R (u, v) shown in FIG. 8 are extracted. A total of 144 reflection characteristics are defined by combinations of 12 φL (φL1 to φL12) and 12 φV (φV1 to φV12).

  When creating the compressed texture data 25, the illustrated tables R1 and R2 can be used as they are for the representative pixels T1 and T2, but the original table Ri (recorded in the original texture data 20) is used for the general pixels Ti. In place of the table), one of the tables R1 and R2 having a high degree of approximation must be substituted. According to the conventionally proposed general compression method, the degree of approximation between the table Ri and the table R2 is higher than the degree of approximation between the table Ri and the table R1, as described above. The most approximate table is the table R2. As a result, in the compressed texture data 25, data indicating that the table R2 of the representative pixel T2 is substituted for the general pixel Ti in the illustrated region C is recorded.

  Actually, considering the direction of the hair row fibers 31, 32, 33 in the regions A, B, C shown in FIG. 10, the reflection characteristics of the region C in which the hair row fibers 33 extending diagonally rightward are planted are completely opposite. It is considered to be preferable to substitute the reflection characteristic of the region B in which the hair row fibers 32 extending in the vertical direction are substituted, instead of substituting the reflection property of the region A in which the hair row fibers 31 extending in the direction are planted. . In addition, in FIG. 11, if the degree of approximation between the tables Ri-R1 and the degree of approximation between the tables Ri-R2 are quantitatively evaluated, the latter will be higher than the former. Therefore, in the conventional general compression method, the representative pixel T2 is selected as the approximate representative pixel for the general pixel Ti.

  However, the inventor of FIG. 10 shows an azimuth angle of the row of fibers 31 planted in the region A in FIG. 10 (an angle indicating the direction when the row of fibers 31 is projected onto the sheet surface of the fiber sheet 30, FIG. When it is shifted by 180 ° (corresponding to an angle of 0 to 360 °), attention is paid to the fact that the hairs are aligned in the same direction as the hair row fibers 33 implanted in the region C. That is, if the inclination angle to the left of the hair row fibers 31 is equal to the inclination angle to the right side of the hair row fibers 33, the hair row fibers 31 with the azimuth angle shifted by 180 ° completely coincide with the hair row fibers 33. It will be. In other words, in FIG. 10, when the illumination light L is irradiated to the position of Ti from the direction of the predetermined azimuth angle φL, the reflection characteristic of the reflected light V emitted in the direction of the predetermined azimuth angle φV is T1 This is equivalent to the reflection characteristic of the reflected light V emitted in the direction of the predetermined azimuth angle (φV + 180 °) when the position is irradiated with the illumination light L from the direction of the predetermined azimuth angle (φL + 180 °).

Therefore, a new table is defined by shifting the azimuth angles φL and φV of the table R1 defined for the representative pixel T1 in the region A by 180 °. The table R1 ^* shown on the left side of FIG. 12 is a table defined in this way. Here, column headings “φL7”, “φL8”,... Of the table R1 ^* are column headings “φL7”, “φL8”,... (Shown as “φL7” in FIG. The row headings “φV7”, “φV8”,... Of the table R1 ^* correspond to the row headings “φV7”, “φV8” of the table R1 shown on the left side of FIG. ,... (In FIG. 11, “φV7” and “φV8” are not shown).

In the two tables R1, R1 ^* , the contents of the rows and columns specified by the same heading are the same. In other words, the table contents are shifted so that the contents of the same row and column with the same heading are the same. For example, the content of the first column indicated by the column heading “φL7” of the table R1 ^* matches the content of the seventh column indicated by the column heading “φL7” of the table R1 (rows are switched). The content of the first row indicated by the row header “φV7” of the table R1 ^* matches the content of the seventh row indicated by the row header “φL7” of the table R1 (columns are switched). More specifically, the tables R1 and R1 ^* are both 12 rows and 12 columns, but the contents of the cell in the first row and the first column (the upper left cell) of the table R1 ^* are the same as those in the table R1. It is equal to the contents of the cell in the seventh row and the seventh column. If such a table R1 ^* is created, the table closest to the table Ri is not the table R2, but the table R1 ^* .

After all, in the example shown in FIG. 10, when selecting the approximate representative pixel for the general pixel Ti, if the conventional method is adopted, the representative pixel T2 is selected as the approximate representative pixel. If the method according to the present invention is adopted, the representative pixel T1 can be selected as an approximate representative pixel by taking into account the condition of “shifting the azimuth angle by 180 °”. In this case, for the general pixel Ti, instead of recording the table Ri, information “representative pixel T1” and information “shift azimuth 180 °” may be recorded. In the actual rendering process, when the general pixel Ti is referred to, the table R1 recorded in the “representative pixel T1” is extracted based on the recorded information, and then the “azimuth angle is shifted by 180 °” The table R1 ^* is created by performing the processing described above, and the reflection characteristics may be determined using the table R1 ^* .

<<< §5. Basic idea of data compression method according to the present invention >>>
As described above, in §4, the representative pixel T1 is set as a general pixel under the condition of “shifting the azimuth angle by 180 °” by utilizing the fact that the inclination directions of the hair row fibers 31 and 33 shown in FIG. An example of successful selection as an approximate representative pixel of Ti is shown. However, actually, such an optimal representative pixel does not always exist. Therefore, in practice, multiple variations of azimuth angle shift amounts (hereinafter referred to as offset amounts) are set, and the table for each representative pixel is shifted by these multiple offset amounts. Is defined as the closest approximation table, and the representative pixel derived from the closest approximation table is selected as the approximate representative pixel. .

  FIG. 13 is a plan view showing the principle of defining a table having a plurality of variations by rotating the contents of the table by shifting the azimuth angle φL of the illumination light L by a predetermined offset amount. If FIG. 13A is an “original table”, FIG. 13B corresponds to a table obtained by shifting φL by 30 °. In this example, the shift for φV is not performed. Here, paying attention to the column headings, in the table of FIG. 13A, the column heading of the first column is “φL1”, and subsequently, “φL2”, “φL3”,. 13 is “φL12”, whereas in the table of FIG. 13B, the column heading of the first column is “φL2”. Hereinafter, “φL3”, “φL4”,. Subsequently, the column heading of the 12th column is “φL1”.

  After all, in the “original table” shown in FIG. 13 (a), the contents of the first column are turned to the twelfth column, the contents of the second column are moved to the first column, If the contents in the third column are rotated to the second column, and each of them is rotated to the left by one column, the table shown in FIG. 13 (b) (a table obtained by shifting φL by an offset amount of 30 °) is obtained. Will be obtained. Similarly, if the contents of the “original table” shown in FIG. 13 (a) are rotated to the left by two columns, the table shown in FIG. 13 (c) (φL is shifted by an offset amount of 60 °). Table). Hereinafter, each table of FIGS. 13D to 13L is a table in which the azimuth angle φL is shifted by an offset amount of 90 ° to 330 ° with reference to the table shown in FIG. 13A.

  In this way, by rotating the azimuth angle φL by an offset amount in increments of 30 °, a table having a total of 12 variations as shown in the figure can be obtained. These 12 tables all contain the same data as the contents, but “in which column each data is included” are different from each other.

In FIG. 13, the azimuth angle φL of the incident illumination light L is rotationally shifted by an offset amount of 30 °, and the azimuth angle φV of the reflected light V to be emitted is not shifted. For each of the tables shown in FIGS. 13 (a) to 13 (l), variations are also created in which the azimuth angle φV of the reflected light V is rotationally shifted by an offset amount of 30 °. For example, in the table of FIG. 13 (a), the content of the first row is turned to the twelfth row, the content of the second row is moved to the first row, and the content of the third row is changed. Turning to the second line and rotating the lines upward by one line respectively, a table obtained by shifting φV by an offset amount of 30 ° is obtained. The table R1 ^* shown on the left side of FIG. 12 is a table obtained by further shifting “φV by 180 °” with respect to the “table obtained by shifting φL by 180 °” shown in FIG. 13 (g). .

As shown in FIGS. 13A to 13L, the azimuth angle φL is rotationally shifted with respect to the “original table” shown in FIG. When 12 variations are created, and each of these 12 variations is further rotated by azimuth angle φV by an offset amount of 30 °, a variation shown in FIG. A total of 144 variations derived from this table are obtained. The table R1 ^* shown on the left side of FIG. 12 is one of the 144 variations.

  The basic idea of the data compression method according to the present invention is that, for example, when two representative pixels T1 and T2 are defined as shown in FIG. 11, the angle parameter φL of the tables R1 and R2 defined for these representative pixels. , ΦV are respectively shifted by various offset amounts to define a table having a plurality of variations, and when finding a table that is closest to the table Ri for the general pixel Ti, Is to be compared. In the case of the simple example shown in FIG. 11, in the conventional method, there are only two candidates for the table Ri and R2, which are the closest to the table Ri, but in the method of the present invention, variations are made by the method described above. As a result, 144 types of tables derived from the table R1, 144 types of tables derived from the table R2, and a total of 288 types of candidates are obtained.

  Of course, when a table closest to the table Ri is selected from the candidates that have caused such variations, in order to designate a table to substitute for the table Ri, which representative pixel is derived from which table It is necessary to identify whether it is a variation. Therefore, in the conventional method, for example, for the general pixel Ti shown in FIG. 11, it is sufficient to record information indicating “approximate representative pixel T2,” but in the method of the present invention, for example, the general pixel Ti shown in FIG. In addition to information indicating “approximate representative pixel T1”, it is also necessary to record information indicating an offset amount of “φL offset amount 180 °, φV offset amount 180 °”. However, since the data capacity of information indicating such an offset amount is extremely small compared to the data capacity of the entire table Ri, even if the method of the present invention is used instead of the conventional method as a texture data compression method, The data compression rate hardly decreases.

  On the other hand, paying attention to the quality of the texture data after compression, it is clear that the quality of the compressed texture data obtained by the method of the present invention is much higher than the quality of the compressed texture data obtained by the conventional method. is there. In the case of the above example, in the conventional method, there are only two types of substitution candidates for the table Ri, the tables R1 and R2, whereas in the method of the present invention, the number of candidates increases dramatically to 288 types. Naturally, the degree of approximation of the table to be substituted is also improved, and significant quality improvement is expected. Therefore, if the method of the present invention is adopted, even if the number of representative pixels is somewhat reduced, it is possible to obtain compressed texture data having a higher quality than the conventional method.

  For example, in the example shown in FIG. 9, four pixels T1 to T4 are determined as representative pixels from a total of 48 pixels, and the remaining 44 general pixels Ti are any of the four tables R1 to R4. However, if the method of the present invention is adopted, even if the number of representative pixels is reduced to three, compressed texture data having higher quality than the conventional method can be obtained. You can expect to get. For example, even if there are only three representative pixels T1 to T3, if 144 variations are created for each of the representative pixel tables R1 to R3, the number of substitution table candidates increases to 432 types. Compared with the three types of candidates in the case of the conventional method, this is a dramatic improvement.

  In the end, if the compression method according to the present invention is adopted, it becomes possible to cope with fewer representative pixels than before, and the overall compression rate can be improved, and compression with higher quality than before is possible. Texture data can be created. In addition, since compression and decompression can be performed by a table content shift operation, high-speed decompression processing is possible.

  11 to 13 show only tables for specific elevation angles θL and θV for convenience of illustration, but in actuality, these tables are defined as one pixel constituting the texture data in the BTF model. It is the partial table which extracted and showed only a part of table which showed the reflected characteristic made. In the case of the BTF model, the original table indicating the reflection characteristics defined for one pixel is a matrix having a size of 72 rows and 72 columns, as in the table R (u, v) shown in FIG. The table shown in FIGS. 11 to 13 is a partial table (12 rows by 12 columns) extracted from only the portion corresponding to θL1 and θV1 (the upper left portion) in the table R (u, v) shown in FIG. Size matrix).

  Therefore, in practice, the “original table” for deriving the variation is not the partial table shown in FIG. 13A but the table R (u, v) as shown in FIG. For example, a table obtained by shifting φL by 30 ° is not a partial table shown in FIG. 13B but a table as shown in FIG. The partial table shown in FIG. 13 (b) corresponds to the table shown in FIG. 14, for example, in which only the portions corresponding to θL1 and θV1 (the upper left portion) are extracted.

  However, in the table shown in FIG. 14, when focusing on the partial table corresponding to “θL2, θV1”, the column heading of the first column is “φL2”, and “φL3”, “φL4”,. Subsequently, the column heading of the twelfth column is “φL1”. Therefore, in the partial table corresponding to “θL2, θV1” in the original table shown in FIG. 8, the contents of the first column are turned to the twelfth column, and the contents of the second column are changed to the first column. , Turn the content in the third column to the second column, and so on, and shift them to the left by one column, thereby causing the partial table corresponding to “θL2, θV1” shown in FIG. Will be obtained. The same applies to other partial tables.

  On the other hand, FIG. 15 is a table in which φV is shifted by 30 ° with the table R (u, v) shown in FIG. 8 as the “original table”. In this table, for each partial table, the content of the first row is turned to the twelfth row, the content of the second row is moved to the first row, and the content of the third row is changed. It can be obtained by turning to the second line and rotating the lines upward by one line.

  After all, when four representative pixels T1 to T4 are defined as in the example shown in FIG. 9, the tables R1 to R4 for each of these representative pixels are actually the sizes (72 shown in FIG. 8). This is a table R (u, v) of a matrix of 72 rows and columns. Therefore, 144 variations of the same size are created for each of the tables R1 to R4. The table shown in FIGS. 14 and 15 is one of the 144 variations. On the other hand, for the general pixel Ti, a table Ri comprising a matrix of 72 rows and 72 columns is prepared in the same way. When four representative pixels are defined as shown in FIG. Therefore, the table closest to this table Ri is selected.

<<< §6. Basic procedure of texture data creation method according to the present invention >>>
Next, the basic procedure of the rendering texture data creation method according to the present invention will be described with reference to the flowchart of FIG. In this basic procedure, step S11 does not necessarily have to be executed using a computer, but each stage of steps S12 to S14 is processing premised on being executed by a computer.

  First, in the original data preparation stage of step S11, it consists of a set of M pixels, and for each individual pixel, both the incident direction or reflection direction of light, or both the incident direction and reflection direction (in this application, “incident direction”). Original texture data in which a table indicating reflection characteristics determined depending on an angle with respect to (and / or a reflection direction) is defined is prepared. In the above example, the texture data 20 shown in FIG. 8 corresponds to the original texture data prepared in step S11.

  In the embodiment shown here, as described above, an example of handling a texture of a BTF model on the assumption that rendering based on the BTF model is performed will be described, but the application of the present invention is limited to the texture of the BTF model. It is not a thing. However, practically, the application to the BTF model is the most preferable form, and the following description will be described on the assumption that this is applied to the texture of the BTF model.

  In the BTF model, as already described with reference to FIG. 4, the sample point Q, the normal n set at the position of the sample point Q, the sample point on the surface of the virtual object to which the texture data is to be mapped, A reference plane S including Q and orthogonal to the normal n and a reference line ζ on the reference plane S passing through the sample point Q are defined. The angle between the illumination light L incident on the sample point Q and the normal line n is θL, the angle between the projection image L ′ of the illumination light L on the reference surface S and the reference line ζ is φL, and the sample point Q When the angle between the reflected light V emerging from the normal line n is θV and the angle between the projected image V ′ of the reflected light V on the reference surface S and the reference line ζ is φV, “θL, φL , ΘV, φV ”, texture data defining a reflection coefficient K (θL, φL, θV, φV) as a table is used. Therefore, the original texture data prepared in step S11 is, for example, data as shown in FIG. The reflection coefficient K can be expressed as a function having six variables K (θL, φL, θV, φV, u, v) by adding the coordinate values u, v of the two-dimensional uv coordinate system as variables. Is as already described.

  The reflection coefficient K (θL, φL, θV, φV) can also be defined as a function expression with θL, φL, θV, φV as variables, but in the present invention, each reflection coefficient K (θL, (φL, θV, φV) are assumed to be given as numerical tables, and are actually defined in the format of a table R (u, v) as illustrated in FIG. In each cell of the table R (u, v), as already described, numerical values Kr, Kg, and Kb indicating the reflectivities of the illumination lights having the wavelengths of the three primary colors RGB are included as the data indicating the reflection characteristics. It is recorded as content. However, in the case of texture data used for rendering processing only for monochrome image processing, it is not necessary to prepare such reflectance for each primary color, and only data showing a single reflectance is recorded. It is enough if you put it.

When the reflection coefficient K is defined as such a numerical table, since continuous amounts cannot be defined for the variables θL, φL, θV, and φV, a discrete angle value within a predetermined range is defined in advance. It is necessary to keep. In the case of the above-described embodiment, with respect to the azimuth angles φL and φV, as shown in FIG. 7, by dividing the angle range of 0 to 360 ° into 12 equal parts, 12 angles 0 °, 30 °, 60 °,. 330 ° is set, but more generally, the angle range of 0 to 360 ° is equally divided by α so that the angle value of φL can be set α and the angle value of φV can be set α. That's fine. Similarly, with respect to the elevation angles θL and θV, as shown in FIG. 6, by dividing the angle range of 0 ° to 90 ° into 6 equal parts, 6 angles 0 °, 15 °, 30 °, 45 °, 60 ° 75 °, but more generally, by dividing the angle range from 0 to 90 ° into β equal parts, θL angle values can be set in β ways and θV angle values can be set in β ways. That's fine. As a result, a table showing a total of α ² × β ² reflection characteristics can be prepared for each pixel. In the embodiment described above, α = 12, β = 6. Therefore, as shown in FIG. 8, α ² × β ² = 5184 reflection characteristics are obtained for one pixel T (u, v). A table R (u, v) shown can be prepared.

  The original texture data to be prepared in step S11 can be artificially created on a computer using CG technology, but a real fiber sheet is photographed using a scanner device or a digital camera. It is also possible to prepare. When creating artificially using the CG technology, prepare the three-dimensional structure data of the virtual fiber sheet (for example, the three-dimensional data representing the state of hairs), and the individual positions of the virtual fiber sheet. The sample point Q is defined as follows, and the intensity of the reflected light V emitted with various angles θV and φV when the angles θL and φL of the incident illumination light L are changed variously may be obtained by simulation.

  On the other hand, when preparing by photographing a real fiber sheet, a plurality of photographed images obtained by photographing from various directions while illuminating illumination light from various directions on a real object The original texture data may be prepared based on the above. Specifically, the illumination environment is adjusted so that the angles θL and φL of the incident illumination light L vary, and the camera position is changed to capture an image in which the angles θV and φV of the reflected light V vary. That's fine. Based on the pixel value of each captured image at the same position on the actual fiber sheet, the “angle-dependent reflection characteristics” of the position can be obtained.

  Now, in the representative pixel determination stage in step S12 in FIG. 16, a process of determining N (N <M) representative pixels from the M pixels constituting the original texture data prepared in step S11 is performed. . The remaining (MN) pixels are general pixels. The example shown in FIG. 9 is a setting example in which M = 48 and N = 4. The representative pixel can be determined at random, or can be extracted and determined regularly to some extent.

  In the case of random determination, random number generation is performed on the computer, a coordinate value (u, v) is selected based on the random number, and the process of determining the pixel corresponding to the coordinate value as the representative pixel is represented by N representatives. This can be continued until a pixel is determined. On the other hand, when regularly extracting and determining, for example, a predetermined rule for extracting representative pixels at predetermined intervals may be determined based on the ratio of M and N. Note that the number N of representative pixels is a parameter that affects the compression rate, and may be appropriately set according to the data capacity required for the finally obtained compressed texture data.

  In the subsequent step S13, an approximate representative pixel selection step is performed in which a representative pixel having a reflection characteristic approximate to its own reflection characteristic is selected as an approximate representative pixel for each general pixel. At this time, as described above, for each representative pixel, a table having a plurality of variations is defined by shifting the angle parameter of the table defined by the original texture data by a predetermined offset amount. A table that is most approximate when all of the tables are compared is determined as the most approximate table, and processing is performed to select a representative pixel from which the most approximate table is derived as an approximate representative pixel.

  Specifically, when a shift that adds or subtracts a predetermined offset amount is performed on each angle value, and the angle value after the shift protrudes from one end of the predetermined range as a result of the addition or subtraction Rotation for counting the amount of protrusion from the other end of the predetermined range may be performed. The shift moving direction is determined depending on whether the offset amount is added or subtracted. FIGS. 13A to 13L show an example of the result of such a rotation shift.

As a general rule, when the azimuth angles φL and φV are set to α by dividing the angle range of 0 to 360 ° by α equally, each unit angle Uφ given by the equation Uφ = 360 ° / α And φL and φV are respectively shifted by α different offset amounts Δφ given by Δφ = i × Uφ (where i = 0, 1, 2,..., Α−1). α It is possible to define a table with ^two combinations of variations. In the embodiment described above, α is set to 12 and the unit angle Uφ is set to 30 °. As a result, for each of φL and φV, 12 variations with offset amounts Δφ = 0 °, 30 °, 60 °,..., 330 ° can be created, and the “original table” for one representative pixel. Therefore, a table having a total of 144 combinations can be derived.

  If there are N representative pixels, the table to be compared will eventually have 144 × N variations. Therefore, in this case, for each of the general pixels Ti, a process that selects the table closest to the table Ri indicating its own reflection characteristic from among the 144 × N variations is performed.

  Various methods are known as methods for quantitatively evaluating the degree of approximation between two tables. For example, for the two tables to be compared, the sum of squares of the differences between the corresponding individual reflectances. May be used as a parameter indicating the degree of approximation, and the table having the minimum value may be determined as the most approximate table.

For example, consider a case where two sets of tables R (u, v) having a size as shown in FIG. In this case, the reflectance recorded in the specific cell in the first table is Kr1, Kg1, Kb1, and the reflectance recorded in the specific cell at the same position in the second table is Kr2, Kg2. , Kb2, the sum of squares of the difference regarding the cell is represented by “(Kr1−Kr2) ² + (Kg1−Kg2) ² + (Kb1−Kb2) ² ”. If the sum of squares of such differences is obtained for all cells (5184 cells in the case of the example in FIG. 8), the sum is used as a parameter indicating the degree of approximation between the first table and the second table. Good. The comparison target with the smallest parameter value is the most approximate table, and the representative pixel derived from the most approximate table is selected as the approximate representative pixel.

In the last step S14, a table indicating its own reflection characteristic is recorded as it is for N representative pixels among the M pixels constituting the original texture data, and (MN) general pixels are recorded. Performs a data compression step of creating compressed texture data by recording data indicating the approximate representative pixel selected in step S13. At this time, for general pixels, data indicating the selected approximate representative pixel and data indicating the offset amount for the table (most approximate table) that causes the selection of the approximate representative pixel are recorded. For example, as in the example shown in FIG. 12, if the most approximate table of the table Ri indicating the reflection characteristics of the general pixel Ti is the table R1 ^* , instead of recording the table Ri for the general pixel Ti. And an offset amount for deriving the data representing the approximate representative pixel T1 and the “original table” for the approximate representative pixel T1 to obtain the most approximate table R1 ^* (in this example, ΔLφ = 180 °, Data indicating ΔVφ = 180 °) are recorded.

  Thus, if the compressed texture data 25 as shown in FIG. 9 is created in step S14, the data structure is as follows. First, for the four representative pixels T1 to T4, the tables R1 to R4 are recorded as they are as in the original texture data. Since these tables have the same size as the table R (u, v) showing the reflection characteristics shown in FIG. 8, the effect of reducing the data capacity is not seen at all for the representative pixels. On the other hand, for the remaining general pixels Ti, such a large-capacity table is not recorded, which is a factor for selecting data representing the approximate representative pixel (for example, T1) and the approximate representative pixel. Since only the data indicating the offset amount (for example, ΔLφ = 180 °, ΔVφ = 180 °) for the most similar table is recorded, the data capacity is greatly reduced.

<<< §7. Basic procedure of rendering method according to the present invention >>
Next, a basic procedure of a rendering method using texture data according to the present invention will be described with reference to the flowchart of FIG. The rendering process described here is basically the same process as the rendering process described in Section 1 with reference to FIG. That is, the purpose is to map the texture data indicating the reflection characteristics of each part onto the surface of the three-dimensional virtual object, and to generate two-dimensional data indicating a projection image obtained when observing the texture data from a predetermined viewpoint. . However, it is assumed that the compressed texture data 25 created by the procedure described in §6 is used as the texture data to be used.

  As described above, the compressed texture data 25 is characterized not only by data indicating the approximate representative pixel but also by deriving the “original table” for the approximate representative pixel for the approximate approximation of the general pixel Ti. Data indicating the offset amount for obtaining the table is recorded.

  Each stage of steps S21 to S27 shown in FIG. 17 is a process executed by a computer. First, in step S21, an object data input step for inputting object data 10 indicating the structure of a three-dimensional virtual object is executed. Subsequently, in step S22, a texture data input step for inputting the compressed texture data 25 is executed. Further, in step S23, a rendering condition setting step is executed for setting a rendering condition for determining an illumination environment, a viewpoint position, a mapping mode, and a projection plane necessary for rendering based on an instruction from the operator. Each of the above stages is a preparation stage for executing the original rendering process.

  Subsequently, the original rendering process shown in steps S24 to S27 is executed by computer computation. First, in step S24, for each sample point Q on the three-dimensional virtual object 10, a corresponding pixel determination step is performed for determining a corresponding pixel in the compressed texture data 25 mapped to the position of the sample point Q. The corresponding pixel mapped to the position of the sample point Q can be determined by referring to the mapping mode set in step S23.

  Next, in step S25, a reflected light calculation stage for calculating the reflected light V from the sample point Q is executed. At this time, the reflection characteristic recorded in the corresponding pixel of the sample point Q is used for the calculation for obtaining the intensity of the reflected light V. That is, when the corresponding pixel is a representative pixel, the intensity calculation of the reflected light V from the sample point Q is performed using the reflection characteristics shown by the table recorded for the representative pixel. On the other hand, when the corresponding pixel is a general pixel, a table recorded for “approximate representative pixel indicated by data recorded for the general pixel” is referred to as “data recorded for the general pixel. The intensity calculation of the reflected light V from the sample point Q is performed using the reflection characteristic indicated by the table obtained by shifting by the “offset amount indicated by”.

  For example, when the corresponding pixel of the sample point Q is the representative pixel T1 shown in FIG. 9, the reflection characteristic indicated by the table R1 recorded for the representative pixel T1 is used to change the sampling point Q from the sample point Q. The intensity calculation of the reflected light V is performed. The table R1 is, for example, a table composed of a matrix of 72 rows and 72 columns as shown in FIG. When the reflected light V from the sample point Q is calculated using the BTF model, the values of θL, φL, θV, and φV are determined based on the object data input in step S21 and the rendering conditions set in step S23. Therefore, the reflectance Kr, Kg, Kb defined in the corresponding cell can be obtained by referring to the table composed of the matrix of 72 rows and 72 columns (the values of θL, φL, θV, φV are Since it is defined discretely, it is possible to perform interpolation if necessary). Based on the reflectances Kr, Kg, and Kb thus obtained, the intensity for each primary color component of the reflected light V is calculated.

  On the other hand, the corresponding pixel of the sample point Q is the general pixel Ti shown in FIG. 9, and data such as “approximate representative pixel: T1”, “offset amount: ΔLφ = 30 °, ΔVφ = 0 °” is recorded. If so, the intensity calculation of the reflected light V from the sample point Q is performed in the following procedure. First, a table R1 recorded for the designated approximate representative pixel T1 is read (as described above, this table R1 is a table composed of a matrix of 72 rows and 72 columns as shown in FIG. 8, for example). Then, a rotation shift corresponding to “offset amount: ΔLφ = 30 °, ΔVφ = 0 °” is executed on this table R1, and for example, a table composed of a matrix of 72 rows and 72 columns as shown in FIG. Get. After that, by referring to this table, specific reflectances Kr, Kg, Kb are obtained, and the intensity for each primary color component of the reflected light V may be calculated.

  As described above, the table content shift operation itself is not a burdensome operation, so even if it is executed in a small-scale system such as a personal computer, the load on the CPU does not increase so much. In practice, it is not necessary to perform a rotation shift on the entire table composed of a matrix of 72 rows and 72 columns as shown in FIG. 8, and to obtain the entire table composed of a matrix of 72 rows and 72 columns as shown in FIG. For example, it is sufficient to perform a rotation shift only for a specific partial table including a cell to be referred to, such as “partial table for θL1 and θV1”.

  Subsequently, in step S26, the processes in steps S24 and S25 described above are repeatedly executed for all necessary sample points. And in the last step S27, based on the reflected light from each sample point Q, the two-dimensional data generation stage which produces | generates the two-dimensional data which show the projection image 15 obtained on the projection plane H is performed.

<<< §8. Basic configuration of apparatus according to the present invention >>
Here, an apparatus for executing the “rendering texture data creation method” shown in the flowchart of FIG. 16 and the “rendering method using texture data” shown in the flowchart of FIG. 17 are shown in the block diagram of FIG. The description will be given with reference. In FIG. 18, the texture data creation apparatus 100 shown in the upper part is an apparatus for executing steps S12, S13, and S14 in FIG. 16, and the rendering apparatus 200 shown in the lower part performs steps S21 to S27 in FIG. It is a device for executing. Each device is a device realized by incorporating a dedicated program into a computer, and the constituent elements shown as individual blocks are realized by a combination of computer hardware and software.

  First, the texture data creation apparatus 100 includes an original data storage unit 110, a representative pixel determination unit 120, a data compression unit 130, and an approximate representative pixel selection unit 140, as illustrated. The texture data creation apparatus 100 is an apparatus that performs processing for creating compressed texture data based on the original texture data before compression, and is preferably configured by a computer system with a certain degree of performance. As described above, since the original texture data before compression has a large capacity of 4 GB, for example, it is difficult to handle it with a general personal computer.

  The original data storage unit 110 is a component having a function of inputting and storing the original texture data prepared in step S11 of FIG. In other words, the original data storage unit 110 is composed of a set of M pixels, and for each individual pixel, an original table in which a table indicating reflection characteristics determined depending on an angle with respect to an incident direction and / or a reflection direction of light is defined. It has a function of inputting texture data and storing it.

  The representative pixel determining unit 120 is a component that determines N (N <M) representative pixels from the M pixels and uses the remaining (MN) pixels as general pixels. The contents are as described in step S12 of FIG.

  Further, the approximate representative pixel selecting unit 140 receives information indicating N representative pixels from the representative pixel determining unit 120, and each of the remaining (MN) general pixels has a reflection characteristic approximate to its own reflection characteristic. A process of selecting a representative pixel having an approximate representative pixel is executed. The details of this process are as described in step S13 of FIG. That is, for each representative pixel, a table having a plurality of variations is obtained by shifting the angle parameter of the table defined by the original texture data stored in the original data storage unit 110 by a predetermined offset amount. Is defined as the closest approximation table when all of these tables are set as comparison targets, and a process of selecting a representative pixel derived from the closest approximation table as an approximate representative pixel is executed.

  The data compression unit 130 directly records a table indicating its own reflection characteristic for each of N representative pixels among the M pixels constituting the original texture data stored in the original data storage unit 110. , (MN) general pixels, together with data indicating the approximate representative pixel selected by the approximate representative pixel selection unit 140, the offset amount for the most approximate table that causes the approximate representative pixel to be selected is set. The process for creating the compressed texture data is executed by recording the indicated data. The details of this process are as described in step S14 in FIG.

  On the other hand, the rendering device 200 includes an object data storage unit 210, a texture data storage unit 220, a rendering condition setting unit 230, a corresponding pixel determination unit 240, a reflected light calculation unit 250, and a two-dimensional data generation unit 260, as illustrated. The compressed texture data created by the texture data creation device 100 is mapped onto the surface of a three-dimensional virtual object, and two-dimensional data indicating a projection image obtained when observing this from a predetermined viewpoint is generated. Fulfills the function. The rendering apparatus 200 does not need to handle large-capacity original texture data, and only needs to have a function that can handle only compressed texture data. Therefore, the rendering apparatus 200 can be configured by a small-scale system such as a general personal computer. is there.

  The object data storage unit 210 performs a function of inputting object data indicating the structure of the three-dimensional virtual object and storing the object data. The texture data storage unit 220 stores the compressed texture data generated by the texture data generation apparatus 100. The rendering condition setting unit 230 has a function of setting a rendering condition for determining an illumination environment, a viewpoint position, a mapping mode, and a projection plane necessary for rendering. These constituent elements are constituent elements that execute the respective steps corresponding to steps S21, S22, and S23 of FIG.

  In addition, the corresponding pixel determination unit 240 has a function of executing a stage corresponding to step S24 in FIG. 17, and refers to the object data, the compressed texture data, and the rendering condition, and the sample point Q on the three-dimensional virtual object. A process of determining a corresponding pixel in the compressed texture data mapped to the position is executed.

  The reflected light calculation unit 250 has a function of executing a step corresponding to step S25 of FIG. 17, and refers to the information indicated by the corresponding pixel in the object data and the compressed texture data, and the rendering condition, from the sample point Q. The intensity of the reflected light is calculated. Specifically, when the corresponding pixel is a representative pixel, the intensity of reflected light from the sample point Q is calculated using the reflection characteristics shown by the table recorded for the representative pixel. When the corresponding pixel is a general pixel, a table recorded for “approximate representative pixel indicated by data recorded for the general pixel” is referred to as “data recorded for the general pixel. The intensity of the reflected light from the sample point Q is calculated using the reflection characteristics shown by the table obtained by shifting by the "offset amount shown by".

  The two-dimensional data generation unit 260 has a function of executing a step corresponding to step S27 in FIG. 17, and generates projection image data based on object data, reflected light intensity, and rendering conditions. That is, based on the reflected light from each sample point Q, two-dimensional data indicating the projection image 15 obtained on the projection plane H is generated, and this is output as projection image data.

<<< §9. Modified example >>>
Finally, some modifications of the basic embodiment described above will be described.
(1) Variations related to elevation angle In the basic embodiments described so far, for example, by shifting the azimuth angles φL and φV related to the “original table” defined for the representative pixel, variations in the table to be compared are changed. It was derived. However, the derivation of the variation is not necessarily performed by shifting the azimuth angles φL and φV, but may be performed by shifting the elevation angles θL and θV.

  FIG. 19 is a table in which θL is shifted by 15 ° with the table R (u, v) shown in FIG. 8 as the “original table”. Here, focusing on a large column heading relating to θL (here, in order to distinguish it from a column relating to φL, the column relating to θL is referred to as a “large column”), in the table shown in FIG. In the table shown in FIG. 19, the first large column is “θL2”, whereas “θL1”, the second large column is “θL2”,. The second large row is “θL3”,..., And the sixth large row is “θL1”. That is, the contents of the first large column corresponding to “θL1” in the original table shown in FIG. 8 are turned to the sixth large column, and the contents of the second large column are moved to the first large column. The table shown in FIG. 19 is obtained by moving and turning the contents of the third large column to the second large column and rotating the left and right by one large column. Here, regarding the contents of the individual partial tables accommodated in each large row, there is no difference between the table shown in FIG. 8 and the table shown in FIG.

  On the other hand, FIG. 20 is a table in which θV is shifted by 15 ° with the table R (u, v) shown in FIG. 8 as the “original table”. Here, focusing on the headline for θV (here, the line for θV is referred to as “large line” to distinguish it from the line for φV), in the table shown in FIG. .theta.V1 ", the second major row is" .theta.V2 ",..., the sixth major row is" .theta.V6 ", whereas the first major row is" .theta.V2 "in the table shown in FIG. The second major line is “θV3”,..., And the sixth major line is “θV1”. That is, the content of the first row corresponding to “θV1” in the original table shown in FIG. 8 is turned to the sixth row, and the content of the second row is turned to the first row. The table shown in FIG. 20 is obtained by moving, turning the content of the third row to the second row, and rotating up by one row.

  The above is an example in which the offset amount of θL or θV is set to 15 ° and rotated, but if the offset amount is set to 30 °, 45 °, 60 °, 75 °, two large rows (two large) Rotation shift of 3 rows (3 rows), 4 rows (4 rows), and 5 rows (5 rows) is possible. When the “original table” is derived by such a shift of the elevation angles θL and θV, a total of 6 × 6 = 36 variations can be obtained.

  However, in practice, a variation based on the shift of the azimuth angles φL and φV is created rather than a variation based on the shift of the elevation angles θL and θV, and the closest approximation table having a higher degree of approximation can be found for each general pixel. The possibility is considered high. This is because, in the case of an actual fiber sheet, as in the example shown in FIG. 10, there are many cases (cases that are rotationally symmetric) having a structure in which the direction of the fur matches when the azimuth is shifted. This is the same not only for a fiber sheet having a row of fibers implanted in a base fabric such as a carpet or carpet, but also for a fiber sheet in which warp and weft are woven (in this case, the offset amount is 90 ° or 180 °). When set to °, the most approximate table is expected to be obtained).

Therefore, in practice, the variation based on the shift of the elevation angles θL and θV is preferably used in combination with the variation based on the shift of the azimuth angles φL and φV. Specifically, a unit angle Uφ given by the equation Uφ = 360 ° / α is defined by dividing the angle range of 0-360 ° by α, and the angle range of 0-90 ° is divided by β. , Uθ = 90 ° / β is defined as a unit angle Uθ (α = 12, β = 6 in the above example). Then, φL and φV are shifted by α offset amounts Δφ given by the equation Δφ = i × Uφ (where i = 0, 1, 2,..., Α−1), respectively, and θL and θV are , Δθ = j × Uθ (where, j = 0,1,2, ..., β -1) becomes a variation of the combination of total α ² × β ² ways which were shifted by each offset amount [Delta] [theta] of beta street given by the formula You can define a table with

(2) Variation by Random Shift In the basic embodiments described so far, the rotation shift is performed when the variation is derived based on the “original table” defined for the representative pixel. However, the table content shift performed in the present invention is not necessarily limited to the rotation shift, and a random shift may be performed. For example, bring the first column to the seventh column, the seventh column to the ninth column, the ninth column to the third column, the second column and the eighth column It's okay to switch to a decent shift such as ... Alternatively, it may be shifted completely at random including the rows and columns. That is, a random shift may be used in which tiles arranged on the two-dimensional matrix are mixed and re-arranged. In such a random shift, a random offset amount is set for each tile.

  In short, a “different table” is created by replacing the arrangement of table contents (individual tiles in the above example) by some method with respect to the “original table” defined for the representative pixel. What is necessary is just to be able to use as a variation. This increases the number of table candidates that can be used as a substitute for the general pixel table, so that it is more likely that a table with a higher degree of approximation can be selected as the closest approximation table, and compression with higher quality is possible. Texture data can be obtained.

  However, when a predetermined table created as a variation is selected as the most approximate table for one general pixel Ti, what kind of shift is performed on the “original table”, the closest approximate table is It is necessary to record information about whether it can be obtained. In other words, the “offset amount with reference to the original position on the table” must be recorded for each content.

  Therefore, in practice, as shown in the basic embodiments described so far, it is preferable to perform the rotation shift. In this rotation shift, the same offset amount is set for all the contents on the table. Therefore, for example, an angle value is simply specified such as “offset amount: ΔLφ = 30 °, ΔVφ = 0 °”. Thus, shift operations for all contents can be defined. In addition, a shift operation by a computer can be executed by an arithmetic processing with a very light burden.

It is a perspective view which shows the concept of the general rendering process using texture data. FIG. 2 is a perspective view showing the simplest method for calculating the intensity of reflected light from a sample point Q in the rendering process shown in FIG. 1. It is a top view which shows the pixel arrangement | sequence of the two-dimensional texture data which consists of a color image. It is a perspective view for demonstrating the principle of a BRDF model and a BTF model. It is a top view which shows the structure of 6-dimensional texture data used with a BTF model. It is a front view which shows an example of the discrete elevation angle definition in a BTF model. It is a front view which shows an example of the discrete azimuth | direction angle definition in a BTF model. It is a top view which shows the structural example of the 6-dimensional texture data 20 defined using the table R (u, v) which shows a reflection characteristic. It is a top view which shows the compression principle of high-dimensional texture data. It is a side view which shows the surface structure of the fiber sheet used as the origin of a texture. It is a top view which shows the approximation condition of the reflective characteristic in high-dimensional texture data. It is another top view which shows the approximation condition of the reflective characteristic in high-dimensional texture data. It is a top view which shows the principle which rotates the content of a table by shifting an azimuth angle only by the predetermined | prescribed offset amount, and defines the table with a some variation. FIG. 9 is a plan view showing a table obtained by shifting φL by 30 ° with respect to the table R (u, v) shown in FIG. 8. FIG. 10 is a plan view showing a table obtained by shifting φV by 30 ° with respect to the table R (u, v) shown in FIG. 8. It is a flowchart which shows the basic procedure of the production method of the texture data for rendering concerning this invention. It is a flowchart which shows the basic procedure of the rendering method using the texture data which concerns on this invention. It is a block diagram which shows the basic composition of the production apparatus and rendering apparatus of the texture data which concern on this invention. FIG. 9 is a plan view showing a table obtained by shifting θL by 15 ° with respect to the table R (u, v) shown in FIG. 8. FIG. 9 is a plan view showing a table obtained by shifting θV by 15 ° with respect to the table R (u, v) shown in FIG. 8.

Explanation of symbols

10: Object data (3D virtual object)
15: Two-dimensional projection image 20: Texture data (original texture data)
25: compressed texture data 30: fiber sheets 31, 32, 33: row of fibers 100: texture data creation device 110: original data storage unit 120: representative pixel determination unit 130: data compression unit 140: approximate representative pixel selection unit 200: Rendering device 210: Object data storage unit 220: Texture data storage unit 230: Rendering condition setting unit 240: Corresponding pixel determination unit 250: Reflected light calculation unit 260: Two-dimensional data generation unit A, B, C: Area on fiber sheet E: viewpoint G: light source H: projection plane I (L): intensity of illumination light I (V): intensity of reflected light K, Kr, Kg, Kb: reflection coefficient K (θL, φL, θV, φV): reflection Coefficients L, L1 to L6: Illumination light L ′: Projected image of illumination light n: Normal line P: Pixels Q and Q (X, Y, Z) on the projection plane: Sample points R1 to R4: Representative pixels Table Ri indicating reflection characteristics: Table R (u, v) indicating reflection characteristics for general pixels: Table S indicating reflection characteristics: Reference planes S11 to S27: Steps T1 to T4 in the flowchart: Representative pixels Ti: General pixels T (u, v): Pixels constituting the texture u, v: Coordinate axes V of the two-dimensional coordinate system defining the texture V: Reflected light V ': Projected image X, Y, Z of reflected light: Three-dimensional coordinate system Each coordinate axis x, y: Each coordinate axis θL of the two-dimensional coordinate system on the projection plane H: Incident angle θL1 to θL6 of the illumination light L: Incident angle θV of the illumination light L: Exit angle θV1 to θV6 of the reflected light V: Reflected light V emission angle ζ: reference line φL: azimuth angle of illumination light L φL1 to φL12: azimuth angle of illumination light L φV: azimuth angle of reflected light V

Claims (14)

An original data preparation stage in which original texture data is prepared which includes a set of M pixels, and for each individual pixel, a table showing a reflection characteristic defined depending on an angle with respect to an incident direction and / or a reflection direction of light is defined. When,
A representative pixel determining step in which N (N <M) representative pixels are defined from the M pixels, and the remaining (MN) pixels are general pixels;
For each individual general pixel, an approximate representative pixel selection step of selecting a representative pixel having a reflection characteristic approximate to its own reflection characteristic as an approximate representative pixel;
Among the M pixels, N representative pixels are recorded with a table indicating their own reflection characteristics, and (MN) general pixels are stored with data indicating the selected approximate representative pixels. A data compression stage for creating the recorded compressed texture data;
In the method of creating rendering texture data having
In the approximate representative pixel selection step, for each representative pixel, a table having a plurality of variations is defined by shifting the angle parameter of the table defined by the original texture data by a predetermined offset amount, When all these tables are compared, the most approximate table is determined as the most approximate table, and the representative pixel derived from the most approximate table is selected as the approximate representative pixel.
In the data compression stage, for the general pixel, compressed texture data including data indicating the selected approximate representative pixel and data indicating the offset amount for the most approximate table that causes the selection of the approximate representative pixel is recorded. A method for creating texture data for rendering, characterized in that it is created. In the creation method of the texture data according to claim 1,
In the original data preparation stage, a table is defined in which angles relating to the incident direction and / or reflection direction of light are expressed by a plurality of angle values obtained by equally dividing a predetermined range of angles,
At the approximate representative pixel selection stage, a shift is performed by adding or subtracting a predetermined offset amount corresponding to an integer multiple of the equally divided angle with respect to each angle value, and as a result of addition or subtraction, the angle value after the shift Is generated from one end of the predetermined range, rotation for counting the amount of protrusion from the other end of the predetermined range is performed. In the creation method of the texture data according to claim 2,
In the original data preparation stage, on the surface of the virtual object to which the texture data is to be mapped, the sample point Q, the normal n set at the position of the sample point Q, and the normal n including the sample point Q An orthogonal reference plane S and a reference line ζ on the reference plane S passing through the sample point Q are defined, and an angle formed by illumination light incident on the sample point Q and the normal n is θL, The angle formed between the projection image of the illumination light on the reference plane S and the reference line ζ is φL, the angle formed between the reflected light emitted from the sample point Q and the normal n is θV, and the reference plane S The reflection coefficient K (θL, φL, θV, φV) given as a function of “θL, φL, θV, φV”, where φV is an angle formed by the projected image of the reflected light onto the reference line ζ The original texture data is defined as a table. How to create for the texture data Daringu. In the creation method of the texture data according to claim 3,
At the original data preparation stage, the angle range of 0 to 360 ° is equally divided by α to set the angle value of φL to α, the angle value of φV to α, and the angle range of 0 to 90 ° to β By equally dividing, θ angle values are set in β ways, θV angle values are set in β ways, and a table showing total α ² × β ² reflection characteristics for each pixel is prepared. How to create texture data for rendering. In the creation method of the texture data according to claim 4,
In the approximate representative pixel selection stage, a unit angle Uφ given by the equation Uφ = 360 ° / α is defined, and φL and φV are defined as Δφ = i × Uφ (where i = 0, 1, 2,..., Α− 1) A method of creating rendering texture data, characterized by defining a table having a total of α ² combinations of variations each shifted by α offset amounts Δφ given by the following equation. In the creation method of the texture data according to claim 4,
In the approximate representative pixel selection stage, a unit angle Uφ given by an equation Uφ = 360 ° / α is defined, a unit angle Uθ given by an equation Uθ = 90 ° / β is defined, φL and φV are defined as Δφ = i × Uφ (where i = 0, 1, 2,..., α−1) are shifted by α different offset amounts Δφ given by the equation, and θL and θV are changed to Δθ = j × Uθ (where j = 0, 1, 2,..., β-1) defining a table having a total of α ² × β ² combinations of variations shifted by β offset amounts Δθ given by the equation A feature of rendering texture data for rendering. In the creation method of the texture data in any one of Claims 1-6,
A method of creating rendering texture data, wherein a table having a numerical value indicating reflectance for each of illumination lights having wavelengths of three primary colors RGB is defined as data indicating reflection characteristics at the original data preparation stage. In the creation method of the texture data according to claim 7,
In the approximate representative pixel selection stage, with respect to a table indicating its own reflection characteristics and a table to be compared, a table that minimizes the sum of squares of the corresponding individual reflectance differences is determined as the most approximate table. To create texture data for rendering. In the creation method of the texture data in any one of Claims 1-8,
In the original data preparation stage, photographing an actual object from various directions while irradiating illumination light from various directions, and preparing original texture data based on the obtained multiple captured images A method of creating texture data for rendering characterized by   A program for causing a computer to execute a representative pixel determination step, an approximate representative pixel selection step, and a data compression step in the texture data creation method according to claim 1.   Compressed texture data created by the texture data creation method according to claim 1. A projection image obtained by mapping the compressed texture data created by the texture data creation method according to claim 1 onto the surface of a three-dimensional virtual object and observing the same from a predetermined viewpoint. A rendering method for generating the two-dimensional data shown,
An object data input step in which a computer inputs object data indicating the structure of the three-dimensional virtual object;
A texture data input step in which the computer inputs the compressed texture data;
A rendering condition setting stage in which a computer sets a rendering condition for determining a lighting environment, a viewpoint position, a mapping mode, and a projection plane necessary for rendering based on an instruction from an operator;
A corresponding pixel determining step in which a computer determines, for each sample point Q on the three-dimensional virtual object, a corresponding pixel in the compressed texture data mapped to the position of the sample point Q;
When the corresponding pixel is a representative pixel, the computer calculates reflected light from the sample point Q using a reflection characteristic indicated by a table recorded for the representative pixel, and the corresponding pixel Is a general pixel, the table recorded for “approximate representative pixel indicated by the data recorded for the general pixel” is displayed as “the data indicated by the data recorded for the general pixel. A reflected light calculation stage for calculating reflected light from the sample point Q using a reflection characteristic indicated by a table obtained by shifting by an “offset amount”;
A computer that generates two-dimensional data indicating a projection image obtained on the projection plane based on reflected light from each sample point Q; and
A rendering method using texture data, characterized by comprising: A set of M pixels. For each individual pixel, original texture data in which a table showing reflection characteristics determined depending on the angle of incidence and / or reflection direction of light is defined and stored. An original data storage unit,
A representative pixel determining unit that determines N (N <M) representative pixels from the M pixels and uses the remaining (MN) pixels as general pixels;
For each of the general pixels, an approximate representative pixel selection unit that selects a representative pixel having a reflection characteristic approximate to its own reflection characteristic as an approximate representative pixel;
Among the M pixels, N representative pixels are recorded with a table indicating their own reflection characteristics, and (MN) general pixels are stored with data indicating the selected approximate representative pixels. A data compression unit for creating the recorded compressed texture data;
In the rendering texture data creating apparatus comprising:
The approximate representative pixel selection unit defines a table having a plurality of variations by shifting the angle parameter of the table defined by the original texture data by a predetermined offset amount for each representative pixel, When all these tables are compared, the most approximate table is determined as the most approximate table, and the representative pixel derived from the most approximate table is selected as the approximate representative pixel.
For the general pixel, the data compression unit stores compressed texture data including data indicating the selected approximate representative pixel and data indicating an offset amount for the most approximate table that is a factor for selecting the approximate representative pixel. An apparatus for creating texture data for rendering, characterized in that it is created. The compressed texture data created by the texture data creation device according to claim 13 is mapped onto the surface of a three-dimensional virtual object, and two-dimensional data indicating a projection image obtained when observing this from a predetermined viewpoint is obtained. A rendering device for generating,
An object data storage unit for inputting and storing object data indicating the structure of the three-dimensional virtual object;
A texture data storage unit for inputting the compressed texture data and storing it;
A rendering condition setting unit for setting a rendering condition for determining a lighting environment, a viewpoint position, a mapping mode, and a projection plane necessary for rendering;
A corresponding pixel determining unit that determines a corresponding pixel in the compressed texture data mapped to the position of the sample point Q on the three-dimensional virtual object;
When the corresponding pixel is a representative pixel, the reflected light from the sample point Q is calculated using the reflection characteristics indicated by the table recorded for the representative pixel, and the corresponding pixel is a general pixel. In the case of, the table recorded for “approximate representative pixel indicated by the data recorded for the general pixel” is used as the “offset amount indicated by the data recorded for the general pixel”. A reflected light calculation unit that calculates reflected light from the sample point Q using the reflection characteristics shown by the table obtained by shifting only
A two-dimensional data generation unit that generates two-dimensional data indicating a projection image obtained on the projection plane based on reflected light from each sample point Q;
A rendering apparatus using texture data.