JP7020075B2 - Image processing equipment and programs - Google Patents

Description

The present invention relates to an image processing apparatus and a program.

Conventionally, a technique for reproducing the real world by physically based rendering by computer graphics (CG) has been developed. Here, physically-based rendering refers to measuring and modeling optical phenomena such as light scattering, reflection, and absorption.

Non-Patent Document 1 describes an example of physically based rendering.

E.A.Khan et al. “Image-Based Material Editing”, Internet (http://citeseerx.ist.psu.edu/viewdoc/download?rep=rep1&type=pdf&doi=10.1.1.87.1905)

In physics-based rendering, it is necessary to calculate using complicated parameters such as albedo (ratio of reflected light to incident light: reflectance), reflectance, roughness, micro facet (normal distribution), but this technique is used as it is. It is difficult to apply to a two-dimensional image.

により光沢を再現する。ここで、
ｆ_{Ｍａｔｕｓｉｋ}：反射特性
Ｌｉ’：入射光強度
ｄ・ｎ：法線方向
である。これにより、３次元物体のＣＧに比べて計算量が削減されるが、反射特性ｆ_{Ｍａｔｕｓｉｋ}は実測のデータを用いる等しているため、十分な改善に至っていない。 Generally, the method of reproducing the texture of a two-dimensional image is as follows.
(1) Statistical calculation such as skewness of histogram is used without considering reflection characteristics. (2) Reflection characteristics are considered, but there are two patterns that do not use reflection characteristics as strictly as CG. Here, the reflection characteristic is specifically a reflectance distribution function (BRDF: (Bidirectional Reflectance Distribution Function)), and when light is incident on the material from a certain direction with respect to a certain point on the reflection surface, each of them. It is a function that expresses how much light is reflected from the material in the direction of. Of these two patterns, in (2), for example

To reproduce the luster. here,
f _Matusik : Reflection characteristic Li': Incident light intensity d · n: Normal direction. As a result, the amount of calculation is reduced as compared with the CG of the three-dimensional object, but the reflection characteristic _fMasik has not been sufficiently improved because the actually measured data is used.

An object of the present invention is to provide a technique capable of expressing a texture more easily than when a reflection characteristic obtained by actual measurement is used while considering the reflection characteristic.

The invention according to claim 1 is the tilt calculated by using a tilt calculation unit that calculates the tilt of the pixel value for each pixel constituting the original image and a reflection characteristic map that defines the reflection characteristics for each tilt. A strong reflection image generation unit that generates a strong reflection image having a relatively strong reflection region by calculating the reflection intensity according to the above, and a composition that combines the original image and the strong reflection image to generate a texture reproduction image. The composite image generation unit includes an image generation unit, a display unit for displaying the texture reproduction image, and a correction unit for gamma-correcting the pixel value of the original image, and the composite image generation unit includes the gamma-corrected original image and the strength. It is an image processing device that synthesizes reflected images .

In the invention according to claim 2, the inclination calculation unit calculates a normal vector as the inclination, and converts a component of the normal vector into an RGB value to generate a normal image from the original image. The image processing device according to claim 1, wherein the strong reflection image generation unit generates the strong reflection image by using the normal image and the reflection characteristic map.

The invention according to claim 3 is the image processing apparatus according to any one of claims 1 and 2, wherein the reflection characteristic map is adjustable by the user.

The invention according to claim 4 is the image processing apparatus according to any one of claims 1 and 2, wherein the reflection characteristic map is selectable from a plurality of reflection characteristic maps by a user.

The invention according to claim 5 is the image processing apparatus according to any one of claims 1 and 2, wherein the reflection characteristic map can be configured by a user by combining a plurality of reflection characteristic maps.

The invention according to claim 6 is the image processing apparatus according to any one of claims 1 and 2, wherein the reflection characteristic map has at least one or more Gaussian distributions.

The invention according to claim 7 is the image processing apparatus according to claim 6 , wherein at least one of the mean value and the dispersion of the Gaussian distribution can be adjusted by the user.

The invention according to claim 8 further includes a filter processing unit for filtering the strong reflection image, and the composite image generation unit synthesizes the original image and the filtered strong reflection image. The image processing apparatus according to any one of 2.

The invention according to claim 9 is the image processing apparatus according to claim 8, wherein the filter processing is an enhancement processing of the strong reflection region.

The invention according to claim 10 is the image processing apparatus according to claim 8, wherein the filter processing is a blurring process of the strong reflection region.

The invention according to claim 11 is the image processing apparatus according to claim 1 , wherein the gamma value of the gamma correction can be adjusted by a user.

The invention according to claim 12 is the image processing apparatus according to claim 1 , wherein the gamma value of the gamma correction is set according to the area of the strong reflection region.

The invention according to claim 13 is the image processing apparatus according to claim 12 , wherein the gamma value of the gamma correction is set so that the original image becomes darker as the area of the strong reflection region becomes larger.

In the invention according to claim 14 , the strong reflection image generation unit uses different reflection characteristic maps to generate different strong reflection images, and the composite image generation unit uses different strong reflection images. The image processing device according to any one of claims 1 and 2, wherein a texture reproduction image is generated, and the display unit displays a different texture reproduction image.

The invention according to claim 15 has a step of inputting an original image into a computer, a step of calculating an inclination of a pixel value for each pixel constituting the original image, and a reflection characteristic defining the reflection characteristic for each inclination. A step of generating a strongly reflected image having a relatively strong reflection region by calculating a reflection intensity according to the inclination calculated using a map, a step of gamma-correcting the pixel value of the original image, and a gamma. It is a program that executes a step of synthesizing the corrected original image and the strongly reflected image to generate a texture reproduction image, and a step of displaying the texture reproduction image on a display unit.

According to the inventions of claims 1, 2 and 15 , the texture of the original image can be expressed more easily than when the reflection characteristics obtained by actual measurement are used while considering the reflection characteristics.

According to the invention of claim 3-5, various textures can be further expressed by the operation of the user.

According to the invention of claim 6, the reflection characteristic map can be further defined easily.

According to the invention of claim 7, various textures can be further expressed by at least one of the average value and the dispersion.

According to the invention of claim 8-10, further, various textures can be expressed by the filtering process.

According to the inventions of claims 11 and 12 , various textures can be further expressed by gamma correction.

According to the thirteenth aspect of the present invention, it is possible to further express a texture with enhanced contrast.

According to the invention of claim 14 , further, different textures can be expressed and displayed.

It is a functional block diagram of an embodiment. It is a block diagram of the configuration of an embodiment. It is a schematic diagram of the pixel tilt direction calculation. It is explanatory drawing which shows an example of the reflection characteristic map. It is a schematic diagram of a strong reflection image generation. It is a schematic diagram of the texture reproduction image generation. It is explanatory drawing which shows the other example of the reflection characteristic map. It is explanatory drawing which shows the further example of the reflection characteristic map. It is explanatory drawing which shows the further example of the reflection characteristic map. It is a functional block diagram of another embodiment. It is a functional block diagram of still another embodiment. It is a functional block diagram of still another embodiment. It is a processing flowchart of still another embodiment.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

<Embodiment 1>
1. 1. Configuration FIG. 1 shows a functional block diagram of an image processing apparatus according to an embodiment. The image processing device of the embodiment functions as a texture expression device for expressing the texture of the original image. The texture expression device includes a pixel tilt direction calculation unit 12, a reflection characteristic generation unit 14, a strong reflection image generation unit 16, a texture reproduction unit 18, and a display unit 20.

The pixel tilt direction calculation unit 12 inputs the original image 10 and calculates the tilt direction for each pixel of the original image to generate a normal image of the original image 10. The pixel tilt direction calculation unit 12 outputs the generated normal image to the strong reflection image generation unit 16.

The reflection characteristic generation unit 14 generates a reflection characteristic map and outputs it to the strong reflection image generation unit 16. The reflection characteristic generation unit 14 may read out and output the reflection characteristic map stored in advance, or may generate and output the reflection characteristic map in response to an operation instruction from the operator.

The strong reflection image generation unit 16 generates a strong reflection image having a relatively strong reflection region by using a normal image and a reflection characteristic map. The strong reflection image generation unit 16 outputs the generated strong reflection image to the texture reproduction unit 18.

The texture reproduction unit 18 functions as a composite image generation unit, generates a texture reproduction image that reproduces the texture of the original image 10 by synthesizing the original image 10 and a strong reflection image, and outputs the texture reproduction image to the display unit 20. The display unit 20 displays the generated texture reproduction image.

Specifically, the functional block shown in FIG. 1 can be realized by a computer.

FIG. 2 shows a specific configuration of the texture expression device. The texture expression device includes a processor 30a, a ROM 30b, a RAM 30c, an input / output interface (I / F) 30d, a communication I / F 30e, a storage device 30f, and a display 30g.

The one or a plurality of processors 30a reads out and executes a processing program stored in the ROM 30b, HDD, SSD, etc., and uses the RAM 30c as a working memory to function in each part of FIG. 1, specifically, a pixel tilt direction calculation unit. 12, the reflection characteristic generation unit 14, the strong reflection image generation unit 16, and the texture reproduction unit 18 are realized. The processor 30a may be composed of a CPU or a GPU. The processes in the processor 30a are listed below.
(1) Acquire (input) the original image 10.
(2) For each pixel constituting the original image 10, the inclination direction of the pixel value is calculated as a normal vector.
(3) A normal image is generated from the original image 10 using the calculated normal vector.
(4) Generate a reflection characteristic map.
(5) The generated reflection characteristic map is applied to the normal image to generate a strong reflection image.
(6) The original image 10 and the strongly reflected image are combined to generate a texture reproduction image.
(7) The generated texture image is displayed on the display 30g.

In the process of (2), the two orthogonal directions of the original image 10 are further set as x and y directions.
(2-1) Calculate the x-direction tilt vector of the pixel value (2-2) Calculate the y-direction tilt vector of the pixel value (2-3) Outer product of the x-direction tilt vector and the y-direction tilt vector Includes the process of calculating the normal vector by calculating.

Further, the processes (4) to (7) are repeatedly executed in response to an operation instruction from the user. That is, the reflection characteristic map was regenerated according to the operation instruction from the user, the regenerated reflection characteristic map was applied to the normal image to regenerate the strong reflection image, and the original image 10 was regenerated. The strong reflection image is combined to regenerate the texture reproduction image, and the regenerated texture image is redisplayed on the display 30g.

The input / output I / F 30d inputs the original image 10 and outputs it to the processor 30a.

The communication I / F30e transmits / receives data to / from an external server via a network. The processor 30a may access an external server to acquire the original image 10.

The storage device 30f is composed of an HDD, an SSD, or the like, and stores one or a plurality of reflection characteristic maps.

The display 30g displays a texture reproduction image. When the reflection characteristic map is regenerated and the texture reproduction image is regenerated according to the operation instruction from the user, a plurality of regenerated texture reproduction images may be displayed in parallel. Further, the texture reproduction image may be displayed in combination with the reflection characteristic map. For example, when the first texture reproduction image is generated using the first reflection characteristic map and the second texture reproduction image is generated using the second reflection characteristic map, it is associated with the first reflection characteristic map. The first texture reproduction image is displayed, and the second texture reproduction image is displayed in association with the second reflection characteristic map. The display 30g may display the original image 10, the normal image, the reflection characteristic map, the strong reflection image, and the texture reproduction image in association with each other.

2. 2. Normal image FIG. 3 schematically shows a pixel tilt direction generation process generated by the pixel tilt direction calculation unit 12, that is, a normal image generation process.

In FIG. 3, two directions orthogonal to each other in the original image 10 are defined as the x direction and the y direction, and the pixel value of each pixel is represented by the height in the z direction orthogonal to the x direction and the y direction. In other words, it is expressed by a height map (High Map) that represents high and low such that black is low and white is high.
Then, as shown in FIG. 3A, when paying attention to a certain pixel 100, the inclination of the pixel value in the x direction in the pixel 100 is calculated, and as shown in FIG. 3B, in the pixel 100. The slope of the pixel value in the y direction is calculated.

Now, let P (x, y) be the pixel value of the pixel 100 of interest, P (x + 1, y) be the pixel value of the pixel adjacent in the x direction, and P (x, y + 1) be the pixel value of the pixel adjacent in the y direction. ), The inclination in the x direction is
Δx (x, y) = P (x + 1, y) -P (x, y)
And the inclination in the y direction is
Δy (x, y) = P (x, y + 1) -P (x, y)
Is.

Next, by calculating the outer product of the x-direction tilt vector 102x whose component is the x-direction tilt Δx and the y-direction tilt vector 102y whose component is the y-direction tilt Δy, the normal vector 102 of the pixel 100 of interest is calculated. Is calculated. FIG. 3 (c) illustrates the normal vector 102 of the pixel 100 of interest.

The above processing is executed for all the pixels constituting the original image 10, and the normal vector is calculated for all the pixels.

After calculating the normal vector for each pixel, the pixel tilt direction calculation unit 12 converts the component of the normal vector for each pixel into an RGB coordinate system to generate a normal image (normal mapping). Normal mapping is known and is a kind of bump mapping technique in 3D CG, which is a method to realize a detailed appearance without using additional polygons, and x, y of the normal vector of a more detailed object. , Z means an RGB image corresponding to the z coordinate.

For example, the x component of the normal vector (x, y, z) is -1 ≦ x ≦ 1, the y component is -1 ≦ y ≦ 1, the z component is 0 ≦ z ≦ 1, and the red component R is 0 to 255. The value of (8 bits) corresponds to -1.0 to 1.0 of the x component of the normal vector, and the value of 0 to 255 of the green component G corresponds to -1.0 to 1. for y of the normal vector. Corresponds to 0, and the value of 0 to 255 of the blue component B is converted corresponding to 0 to 1.0 of the z component of the normal vector. As a result, when the pixel 100 of interest in the original image 10 is black and the pixels around it are also black, the normal vector becomes (0,0,1), and when this is converted to RBG (R, G, B). ) = (128, 128, 255). This corresponds to bluish purple. When normal mapping is performed, the unevenness becomes deeper as the amount of change in color in the original image 10 increases, and the unevenness becomes finer as the changing distance becomes shorter.

3. 3. Reflection characteristic map FIG. 4 shows an example of the reflection characteristic map 200 generated by the reflection characteristic generation unit 14. The reflection distribution differs depending on the material of the object (surface unevenness, etc.), but the light reflected on the surface of the object can be simply expressed by diffused light and specularly reflected light. The reflection characteristic generation unit 14 generates a reflection characteristic map using, for example, one or a plurality of Gaussian distributions.

In the figure, the reflection characteristic map 200 is represented as a two-dimensional map. The horizontal direction indicates an angle of 0 ° to 180 ° in the x direction, the vertical direction indicates an angle of 0 ° to 180 ° in the y direction, and the reflection intensity for each (x direction angle, y direction angle) is shown in different colors. In the figure, it is shown that the reflection intensity increases as the angle in the x direction approaches 90 ° and as the angle in the y direction approaches 90 °. By referring to the reflection characteristic map 200 shown in FIG. 4 from the angles in the x direction and the y direction of the normal vector of the normal image, the pixel value corresponding to the normal vector can be obtained. The reflection characteristic map 200 is not obtained by actual measurement and can be set arbitrarily.

One or a plurality of reflection characteristic maps 200 are stored in the storage device 30f in advance, and the user can arbitrarily select these reflection characteristic maps 200. For example, the processor 30a displays one or more reflection characteristic maps 200 stored in the storage device 30f on the display 30g and makes them available to the user for selection. When the desired reflection characteristic map is selected by the user instructing the operation, the processor 30a converts the normal vector into the corresponding pixel value using the selected reflection characteristic map 200.

4. Strong reflection image The strong reflection image generation unit 16 generates a strong reflection image from the normal image and the reflection characteristic map 200. Specifically, as described above, the processor 30a converts the normal vector of the normal image into a normal vector by referring to the reflection characteristic map 200 shown in FIG. 4 from the x-direction and y-direction angles of the normal vector. Generate the corresponding pixel value.

FIG. 5 shows an example of a strong reflection image generated from the normal image using the reflection characteristic map 200 shown in FIG. FIG. 5A is a normal image, and FIG. 5B is an image converted by applying the reflection characteristic map 200. The reflection characteristic map 200 gives a relatively strong reflection intensity for one normal vector and a relatively weak reflection intensity for another normal vector. As a result, the reflection intensity becomes stronger for a specific normal vector portion in the normal image. In FIG. 5B, regions 40 and 42 are shown by solid lines in some of the strongly reflecting portions.

5. Texture reproduction The texture reproduction unit 18 generates a texture reproduction image by synthesizing the original image 10 and the strongly reflected image. Specifically, the processor 30a stores the original image 10 in the RAM 30c, stores the generated strong reflection image in the RAM 30c, reads these images from the RAM 30c, and synthesizes them to generate a texture reproduction image.

FIG. 6 shows an example of a texture reproduction image. FIG. 6A is an original image 10, and FIG. 6B is a strong reflection image. The texture reproduction unit 18 generates the texture reproduction image shown in FIG. 6 (c) by synthesizing the original image 10 and the strongly reflected image. That is, the texture reproduction unit 18 reproduces the texture of the original image 10 by adding the strong reflection region generated by the reflection characteristic map 200 to the original image 10.

It should be noted that in the embodiment, the strong reflection image is generated from the normal image of the original image 10 and the reflection characteristic map 200. The color change is expressed as unevenness by the normal image, and the strong reflection region is extracted from the unevenness by using the reflection characteristic map 200 to give a texture to the two-dimensional original image 10.

Further, the strongly reflected image may be displayed as it is as a reproduced image on the display unit without being combined with the two-dimensional original image 10. The user may be able to select whether to display the original image 10 and the strongly reflected image in combination or to display the strongly reflected image as it is.

Since the strong reflection region is generated from the normal image of the original image 10 using the reflection characteristic map 200, the strong reflection region also changes by changing the reflection characteristic map 200, and as a result, the texture of the texture reproduction image also changes. Can be.

FIG. 7 shows an example of a plurality of reflection characteristic maps 202, 204, 206. As shown in FIG. 4, the reflection characteristic maps 202, 204, and 206 are basically two-dimensional maps of angles in the x and y directions, but here, they are simply shown as one-dimensional maps of the angles in the x direction. A two-dimensional map can be easily generated from a one-dimensional map.

FIG. 7A shows a reflection characteristic map 202 of a Gaussian distribution that is relatively flat with respect to an angle in the x direction. In the figure, the horizontal axis is an elevation angle of 0 ° to 180 °, and the vertical axis is the reflectance. Since the reflection characteristics are relatively flat, the strong reflection region is correspondingly slowed down.

FIG. 7B shows a reflection characteristic map 204 of a relatively steep Gaussian distribution. Since the reflection characteristics are relatively steep, a specific strong reflection region is emphasized accordingly.

7 (c) shows a reflection characteristic map 206 which is a combination of the reflection characteristic map 202 shown in FIG. 7 (a) and the reflection characteristic map 204 shown in FIG. 7 (b). Therefore, a strong reflection region is added to the original image 10 at the reflectance shown in the reflection characteristic map 202 when the angle in the x direction is near 0 ° or 180 °, and is shown in the reflection characteristic map 204 when the angle in the x direction is near 90 °. A strong reflection region is added to the original image 10 by the reflectance.

The user can select an arbitrary reflection characteristic map from the plurality of reflection characteristic maps 202, 204, 206 shown in FIGS. 7 (a), 7 (b), and (c) to generate a texture reproduction image. Since the Gaussian distribution is defined by the mean value μ and the variance σ ² , the user changes the reflection characteristic map by changing at least one of the mean value μ and the variance σ ² , and obtains a desired texture reproduction image. Can be generated.

For example, the user sequentially changes the reflection characteristic map by sequentially changing the variance σ ² in the order of σ ₁ ² → σ ₂ ² → σ ₃ ² (however, σ ₁ ² > σ ₂ ² > σ ₃ ² ). Texture reproduction The image is changed sequentially. The user sequentially changes the texture reproduction image, and when the desired texture reproduction image is obtained, the user selects the texture reproduction image according to the operation instruction. The processor 30a stores the texture reproduction image selected by the user in the storage device 30f together with the reflection characteristic map at that time.

The processor 30a communicates the reflection characteristic map selected by the user, the reflection characteristic map and the texture reproduction image selected by the user, or the original image 10 and the reflection characteristic map and the texture reproduction image through the communication I / F30e and the network. The external server may be transmitted to an external server via the above, and the external server may store a reflection characteristic map, a texture reproduction image, or an original image transmitted from the texture reproduction device.

As described above, in the embodiment, instead of using the reflection characteristics obtained by actual measurement, a predetermined reflection characteristic map such as a Gaussian distribution is used, and the reflection characteristic map is appropriately changed to express a desired texture. , The texture of the original image 10 can be expressed more easily.

<Embodiment 2>
FIG. 8 shows a functional block diagram of the texture expression device according to the present embodiment. The difference from FIG. 1 is that the strong reflection region emphasizing portion 22 is added. The strong reflection region enhancement unit 22 is realized by executing a processing program by one or a plurality of processors 30a, but may be configured by a hardware circuit.

The strong reflection image generation unit 16 generates a strong reflection image from the normal image and the reflection characteristic map as in the first embodiment. The strong reflection image generation unit 16 outputs the generated strong reflection image to the strong reflection region enhancement unit 22.

The strong reflection region enhancement unit 22 filters the strong reflection image to emphasize the edges of the strong reflection region. Alternatively, the strong reflection region enhancement unit 22 may filter the strong reflection image to blur the strong reflection region. The strong reflection region enhancement unit 22 outputs the filtered strong reflection image to the texture reproduction unit 18.

According to the present embodiment, a further different texture (for example, a smooth texture) can be expressed by emphasizing or blurring the edge of the strong reflection region.

<Embodiment 3>
FIG. 9 shows a functional block diagram of the texture expression device according to the present embodiment. The difference from FIG. 1 is that the pixel value correction unit 24 is added. The pixel value correction unit 24 is realized by executing a processing program by one or a plurality of processors 30a, but may be configured by a hardware circuit.

The pixel value correction unit 24 filters the original image 10 and outputs it to the texture reproduction unit 18. For example, the pixel value correction unit 24 darkens the pixel value of the original image 10 by gamma correction. The pixel value correction unit 24 outputs the filtered original image 10 to the texture reproduction unit 18.

The texture reproduction unit 18 generates a texture image by synthesizing the filtered original image 10 and the strongly reflected image. By darkening the pixel value of the original image 10, the contrast with the strong reflection region is increased, and a different texture (texture becomes stronger) can be expressed.

The pixel value correction unit 24 may acquire the strong reflection image generated by the strong reflection image generation unit 16 and change the amount of darkening of the original image 10 according to the area of the strong reflection image. For example, the larger the area of the strongly reflected image, the darker the original image 10. Specifically, the input value IN and the correction value OUT in the pixel value correction unit 24 are set to
OUT = IN ^γ
Then, the value of γ is set so that the larger the area of the strongly reflected image, the larger the value. As a result, the contrast with the strong reflection region is further increased, and the texture is further strengthened.

<Embodiment 4>
The first embodiment, the second embodiment, and the third embodiment may be combined.

FIG. 10 shows a functional block diagram of the texture reproduction device according to the present embodiment. The strong reflection region enhancement unit 22 in the second embodiment and the pixel value correction unit 24 in the third embodiment are added to the functional block diagram shown in FIG. The user can generate various texture reproduction images by adjusting at least one of the reflection characteristic map, the filter processing in the strong reflection region enhancement unit 22, and the filter processing in the pixel value correction unit 24.

FIG. 11 shows a processing flowchart in the configuration of FIG.

First, the pixel tilt direction calculation unit 12 acquires the original image 10 (S101).

Next, the pixel tilt direction calculation unit 12 calculates a normal vector for each pixel of the original image (S102). Specifically, the slope vector in the x direction and the slope vector in the y direction are calculated, and the normal vector is calculated by calculating the outer product of these.

Next, the pixel tilt direction calculation unit 12 generates a normal image from the normal vector (S103). That is, the components (x, y, z) of the normal vector for each pixel are converted into RGB values. The pixel tilt direction calculation unit 12 may display the generated normal image on the display unit 20.

Next, the reflection characteristic generation unit 14 generates a reflection characteristic map (S104). As the reflection characteristic map, the reflection characteristic map stored in advance in the storage device 30f may be used as it is, or the reflection characteristic map stored in advance may be changed and generated according to an operation instruction from the user. An example of a change in the reflection characteristic map is a combination of a plurality of reflection characteristic maps as shown in FIG. 7C. Another example is the adjustment of at least one of the mean value μ and the variance σ ² of the Gaussian distribution by the user. A plurality of reflection characteristic map candidates are stored in the external server, and the texture reproduction device accesses the external server, downloads these reflection characteristic map candidates from the external server, displays them on the display unit 20, and the user desires them. The reflection characteristic map may be generated by selecting the reflection characteristic map of.

Next, the strong reflection image generation unit 16 generates a strong reflection image from the normal image and the reflection characteristic map (S105). The strong reflection image generation unit 16 may display the generated strong reflection image on the display unit 20.

Next, the strong reflection region enhancing unit 22 applies a filter treatment to the strong reflection region to emphasize the strong reflection region in a desired manner (S106). The user can select what kind of filtering should be applied. The strong reflection region enhancement unit 22 may be configured to display various filters on the display unit 20 so that the user can select a desired filter process.

Next, the pixel value correction unit 24 gamma-corrects the pixel value of the original image (S107). The user can select the gamma (γ) value at the time of gamma correction. Further, the pixel value correction unit 24 may set the gamma value according to the area of the strong reflection region in the strong reflection image. The pixel value correction unit 24 may display the original image 10 with the pixel value corrected on the display unit 20.

Next, the texture reproduction unit 18 generates a texture reproduction image by synthesizing the original image 10 whose pixel value has been corrected and the strong reflection image in which the strong reflection region is emphasized (S108). The texture reproduction unit 18 displays the generated texture reproduction image on the display unit 20 (S109).

Next, the texture reproduction device determines whether or not the generated texture reproduction image satisfies the requirement (S110). Specifically, it is determined whether or not the processor 30a satisfies the request in response to the input from the user. When the user is satisfied with the texture reproduction image displayed on the display unit 20, the user inputs, for example, YES. On the other hand, if the user is not satisfied with the texture reproduction image, the user inputs to that effect, for example, NO. If the requirement is not satisfied, the processing after S104 is repeated, and at least one of the reflection characteristic map, the filter processing of the strong reflection region, and the pixel value correction of the original image is changed to regenerate the texture reproduction image and display it on the display unit 20. indicate.

In the processing flowchart of FIG. 11, when the processing of S106 and S107 is skipped, it corresponds to the first embodiment, when the processing of S107 is skipped, it corresponds to the second embodiment, and when the processing of S106 is skipped, it corresponds to the third embodiment. Corresponds to.

Although the embodiments of the present invention have been described above, the present invention is not limited to these embodiments and can be modified in various ways.

For example, in the embodiment, the Gaussian distribution is exemplified as the reflection characteristic map, but the present invention is not limited to this, and the reflection characteristic map can be generated by using an arbitrary function or table.

Further, in the embodiment, the inclination vector 102x in the x direction and the direction vector 102y in the y direction of the pixel 100 of interest shown in FIG. 3 are used.
Δx = P (x + 1, y) -P (x, y)
Δy = P (x, y + 1) -P (x, y)
However, as the inclination in the x direction, the average of the inclinations in the x direction of the pixels adjacent to the y direction is Δx = {Δx (x, y-1) + Δx (x, y) + Δx (x, y + 1)} / 3
As the inclination in the y direction, the average of the inclinations of the pixels adjacent to the x direction in the y direction is Δy = {Δy (x-1, y) + Δy (x, y) + Δy (x + 1, y)} / 3
It may be calculated as. here,
Δx (x, y-1): Inclination in the x direction of the pixel (x, y-1) adjacent to the pixel of interest (x, y) Δx (x, y): Inclination in the x direction of the pixel of interest (x, y) ( Before average calculation)
Δx (x, y + 1): Inclination in the x direction of the pixel (x, y + 1) adjacent to the pixel of interest (x, y) Δy (x-1, y): Pixel (x-) adjacent to the pixel of interest (x, y) 1, y) y-direction inclination Δy (x, y): y-direction inclination of the pixel of interest (x, y) (before average calculation)
Δy (x + 1, y): The inclination in the y direction of the pixel (x + 1, y) adjacent to the pixel of interest (x, y).

Further, in the embodiment, the reflection characteristic map 200 shown in FIG. 4 or the reflection characteristic maps 202, 204, 206 shown in FIGS. 7A, 7B, and 7C are shown as the reflection characteristic maps, but the reflection characteristic map is for each pixel. It suffices to specify the inclination of the pixel value, that is, the relationship between the direction of the normal vector and the reflection intensity, and may be in the function format or the table format.

10 original image, 12 pixel tilt direction calculation unit, 14 reflection characteristic generation unit, 16 strong reflection image generation unit, 18 texture reproduction unit, 20 display unit, 102 normal vector, 200, 202, 204, 206 reflection characteristic map.

Claims (15)

A tilt calculation unit that calculates the tilt of the pixel value for each pixel that constitutes the original image,
A strong reflection image generation unit that generates a strong reflection image having a relatively strong reflection region by calculating the reflection intensity according to the calculated inclination using the reflection characteristic map that defines the reflection characteristic for each inclination. When,
A composite image generation unit that synthesizes the original image and the strong reflection image to generate a texture reproduction image,
A display unit that displays the texture reproduction image and
A correction unit that gamma-corrects the pixel value of the original image,
The composite image generation unit synthesizes the gamma-corrected original image and the strong reflection image.
Image processing device. The slope calculation unit calculates a normal vector as the slope and converts a component of the normal vector into an RGB value to generate a normal image from the original image.
The image processing apparatus according to claim 1, wherein the strong reflection image generation unit generates the strong reflection image by using the normal image and the reflection characteristic map. The image processing apparatus according to any one of claims 1 and 2, wherein the reflection characteristic map is adjustable by the user. The image processing apparatus according to any one of claims 1 and 2, wherein the reflection characteristic map can be selected by the user from a plurality of reflection characteristic maps. The image processing apparatus according to any one of claims 1 and 2, wherein the reflection characteristic map can be configured by a user by combining a plurality of reflection characteristic maps. The image processing apparatus according to any one of claims 1 and 2, wherein the reflection characteristic map is at least one Gaussian distribution. The image processing apparatus according to claim 6 , wherein at least one of the mean value and the variance of the Gaussian distribution can be adjusted by the user. Further, a filter processing unit for filtering the strongly reflected image is provided.
The image processing apparatus according to any one of claims 1 and 2, wherein the composite image generation unit synthesizes the original image and the filtered strong reflection image. The image processing apparatus according to claim 8, wherein the filter processing is an enhancement processing of the strong reflection region. The image processing apparatus according to claim 8, wherein the filter processing is a blurring process of the strong reflection region. The image processing apparatus according to claim 1 , wherein the gamma value of the gamma correction can be adjusted by the user. The image processing apparatus according to claim 1 , wherein the gamma value of the gamma correction is set according to the area of the strong reflection region. The image processing apparatus according to claim 12 , wherein the gamma value of the gamma correction is set so that the original image becomes darker as the area of the strong reflection region becomes larger. The strong reflection image generation unit generates different strong reflection images using different reflection characteristic maps.
The composite image generation unit generates different texture reproduction images using the different strong reflection images.
The image processing apparatus according to any one of claims 1 and 2, wherein the display unit displays different texture reproduction images. On the computer
Steps to enter the original image and
The step of calculating the slope of the pixel value for each pixel constituting the original image, and
A step of generating a strong reflection image having a relatively strong reflection region by calculating the reflection intensity according to the inclination calculated by using the reflection characteristic map that defines the reflection characteristic for each inclination.
The step of gamma-correcting the pixel value of the original image,
A step of synthesizing the gamma-corrected original image and the strong reflection image to generate a texture reproduction image, and
The step of displaying the texture reproduction image on the display unit and
A program to execute.

Images