Information processing device, information processing method, and computer-readable non-transitory storage medium

Abstract

This information processing device includes a rendering unit and a detail restoration processing unit. The rendering unit performs rendering processing on a low-detail 3D model in which a detailed structure is reduced. The detail restoration processing unit generates a restored image in which the detailed structure is restored from a low-detail 2D image obtained by the rendering processing.

Description

Information processing device, information processing method, and computer-readable non-transitory storage medium

[0001] The present invention relates to an information processing device, an information processing method, and a computer-readable non-transitory storage medium.

[0002] 3DCG rendering involves sophisticated simulation of light and shadows, such as ray tracing. While this type of simulation process produces high-quality rendered images, it requires a large amount of calculations, making real-time processing difficult even with the latest GPUs (Graphics Processing Units).

[0003] Patent No. 4580938

[0004] LOD (Level of Detail) is known as a technique for reducing the amount of calculation. This technique increases calculation speed by omitting the detailed structure of distant objects. Because these are distant objects, the degradation in image quality caused by omitting the detailed structure is difficult to perceive. "Image quality" refers to the visual quality based on the fidelity (accuracy) of the detailed structure. However, this technique cannot be used in situations where the scene is entirely composed of close-up objects. Because such situations are frequent, there are limited scenes in which LOD technology can be applied without degrading image quality.

[0005] Therefore, the present disclosure proposes an information processing device, an information processing method, and a computer-readable non-transitory storage medium that can more effectively reduce the amount of rendering calculation.

[0006] According to the present disclosure, there is provided an information processing device including a rendering unit that performs a rendering process on a low-detail 3D model in which detailed structures have been reduced, and a detail restoration processing unit that generates a restored image in which the detailed structures have been restored from a low-detail 2D image obtained by the rendering process. The present disclosure also provides an information processing method in which the information processing of the information processing device is executed by a computer, and a computer-readable non-transitory storage medium that stores a program that causes a computer to realize the information processing of the information processing device.

[0007] 1 is a diagram illustrating a rendering technique using LOD technology. FIG. 1 is a diagram illustrating an example of a 3D model prepared for each level of detail. FIG. 2 is a diagram illustrating a conventional rendering technique using LOD technology. FIG. 3 is a diagram illustrating an example of setting a level of detail. FIG. 4 is a diagram illustrating an overview of the rendering technique of the present disclosure. FIG. 5 is a diagram illustrating an example of learning of an inference model that performs detail restoration processing. FIG. 6 is a diagram illustrating an example of learning of an inference model that performs detail restoration processing. FIG. 7 is a diagram illustrating an example of a system configuration of an information processing device used in the rendering phase. FIG. 8 is a diagram illustrating an example of a system configuration of an information processing device used in the level of detail setting phase. FIG. 9 is a diagram illustrating an example of a processing flow of the level of detail setting phase. FIG. 10 is a diagram illustrating an example of level of detail settings before and after update. FIG. 11 is a diagram illustrating an example of a system configuration of an information processing device used in the rendering phase. FIG. 12 is a diagram illustrating an example of an image quality evaluation index value table. FIG. 13 is a diagram illustrating an example of partial table data. FIG. 14 is a diagram illustrating an example of a processing flow of the rendering phase. FIG. 15 is a diagram illustrating an example of a system configuration of an information processing device used in the image quality evaluation index value measurement phase. FIG. 16 is a diagram illustrating an example of a processing flow of the image quality evaluation index value measurement phase.

[0008] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. In the following embodiments, the same components are designated by the same reference numerals, and redundant description will be omitted.

[0009] The description will be given in the following order: [1. Rendering method using LOD technology and its problems] [2. Overview of the rendering method of the present disclosure] [3. Example of system configuration in image quality priority mode] [3-1. Rendering phase] [3-2. Level of detail setting phase] [3-3. Effects] [4. Example of system configuration in speed priority mode] [4-1. Rendering phase] [4-2. Image quality evaluation index value measurement phase] [4-3. Effects] [5. Example of hardware configuration]

[0010] 1. Rendering Method Using LOD Technology and Its Issues FIG. 1 is a diagram for explaining a rendering method using LOD technology.

[0011] LOD technology is a technology that appropriately controls the number of polygons (level of detail) of an object OB according to its distance from the camera (rendering viewpoint PV). LOD technology improves performance by displaying the optimal amount of data while adjusting how it appears on the screen. LOD technology is used in real-time rendering of games, etc.

[0012] For example, an object OB close to the camera Ｎ If you reduce the number of polygons, the shape will look rough. On the other hand, if the object OB is far from the camera, Ｆ The detailed structure of the object OB is difficult for the user to see, so even if the level of detail is reduced, the roughness of the shape is difficult to recognize. Ｆ The level of detail is the close-up object OB Ｎ This speeds up the rendering process without significantly compromising the appearance.

[0013] FIG. 2 is a diagram showing an example of 3D models prepared according to the level of detail.

[0014] A plurality of 3D models with different levels of detail are prepared for each object OB to be rendered. In the example of FIG. 2, four levels of detail, from "Lv0" to "Lv3," are shown. The smaller the number following "Lv," the higher the level of detail. For example, the number of polygons for "Lv0," "Lv1," "Lv2," and "Lv3" are 4096, 1024, 256, and 64, respectively. Although FIG. 2 shows an example with four levels of detail, the number of levels is not limited to four.

[0015] FIG. 3 is a diagram illustrating a conventional rendering method using the LOD technique.

[0016] In a conventional rendering method, first, a viewpoint setting unit 12 sets a rendering viewpoint PV. The rendering viewpoint PV refers to a viewpoint for converting an object OB into a 2D image by rendering processing. A distance calculation unit 13 calculates a distance DS from the rendering viewpoint PV to the object OB.

[0017] The detail level setting unit 15 sets a level of detail LV according to the distance DS based on a detail level setting table LT. The detail level table setting unit 14 includes a detail level setting table LT that defines an appropriate level of detail. For example, the detail level setting table LT defines, for each distance DS, the minimum level of detail LV that provides acceptable image quality on the screen as the appropriate level of detail. The detail level setting unit 15 acquires a 3D model MD of the object OB having an appropriate level of detail from the database unit 11 and outputs it to the rendering unit 16. The rendering unit 16 performs rendering processing on the acquired 3D model MD to generate a 2D image IM having a detailed structure according to the appropriate level of detail.

[0018] 4 is a diagram showing an example of setting the level of detail LV. In the example of FIG. 4, when the distance DS is less than 3, the level of detail LV is set to "Lv0." When the distance DS is equal to or greater than 3 and less than 10, the level of detail LV is set to "Lv1." When the distance DS is equal to or greater than 10 and less than 30, the level of detail LV is set to "Lv2." When the distance DS is 30 or greater, the level of detail LV is set to "Lv3." Note that the value of the distance DS is normalized by a normalization coefficient defined for each scene.

[0019] 2. Overview of the Rendering Method of the Present Disclosure FIG. 5 is a diagram illustrating an overview of the rendering method of the present disclosure.

[0020] In the conventional rendering method described above, the object OB in the distant view Ｆ A low-detail 3D model (MD) is applied to the scene. As a result, the amount of rendering calculations cannot be sufficiently reduced in scenes with little distant view. Because such situations often occur, the scenes to which LOD technology can be applied without degrading image quality are limited. This disclosure has been made in light of these issues. In this disclosure, the following method is adopted to effectively reduce the amount of rendering calculations in various scenes.

[0021] First, a low-detail 3D model with reduced detailed structure is acquired from the database unit 11. The low-detail 3D model with reduced detailed structure means a rough 3D model MD with a reduced number of polygons and a simplified detailed structure. The rendering unit 16 performs rendering processing on the low-detail 3D model to generate a 2D image IM (low-detail 2D image IM) of the object OB with a rough shape. Ｌ ) to generate the

[0022] The detail restoration processing unit 17 generates a low-detail 2D image IM Ｌ The detailed shape of the object OB is restored (detail restoration process) to the low-detail 2D image IM Ｌ A 2D image IM (restored image RI) of the object OB with its detailed structure restored is generated from the 3D model MD. The restored image RI is a highly detailed 2D image IM with the same degree of reproducibility of the detailed shape as when a 3D model MD with a large number of polygons is used, due to the detailed structure being restored. Ｈ This becomes:

[0023] The detailed restoration process is realized by AI learning-based image signal processing using, for example, a convolutional neural network (CNN). Examples of such image signal processing include "generative adversarial networks" (GANs) and "transformers."

[0024] In this rendering method, rendering is performed on a coarse 3D model MD with reduced detailed structures. This results in a significant reduction in the amount of rendering calculations. Although processing is required to restore the detailed shape after rendering, the detail restoration process is performed on a 2D image, so the amount of calculations is smaller than the calculations performed on the 3D model MD. This method can also be applied to foreground objects OB, reducing the amount of rendering calculations in a variety of scenes. This results in a significant reduction in the amount of rendering calculations compared to conventional methods.

[0025] The detail restoration processing unit 17 does not perform detail restoration processing on a 2D image IM whose detailed shape is not omitted. For example, when a 3D model MD with the highest level of detail LV is selected as the rendering target, the detail restoration processing unit 17 outputs the 2D image IM acquired from the rendering unit 16 as a restored image RI without performing the level of detail restoration processing.

[0026] 6 and 7 are diagrams illustrating an example of learning an inference model that performs detailed reconstruction processing.

[0027] The inference model is a highly detailed 2D image IM generated from 3D models MD with different levels of detail LV. Ｈ and low-detail 2D images IM Ｌ are used as the teacher image and the student image, respectively. Ｈ is a highly detailed 3D model MD in which the detailed structure has not been reduced or the degree of reduction in the detailed structure is small. Ｈ The low-detail 2D image IM is obtained by rendering the Ｌ Highly detailed 3D model MD Ｈ Low-detail 3D model MD, which has a greater reduction in the degree of detailed structure than Ｌ is obtained by rendering.

[0028] The quality of the restored image RI is determined by the quality of the teacher image. Ｈ It is preferable that the 3D model MD with the highest level of detail is the 3D model MD with the highest level of detail. In the example of FIG. 7, the 3D model MD with "Level 0" corresponds to this. Therefore, the 3D model MD with "Level 0" is the 3D model MD for the teacher image (high-detail 3D model MD Ｈ ) and the 3D model MDs of "Level 1" to "Level 3" are used as 3D model MDs for student images (low-detail 3D model MDs Ｌ ) is used as

[0029] A specific example of an information processing device for implementing the rendering method of the present disclosure will be described below. The rendering process of the present disclosure has an image quality priority mode and a speed priority mode. The image quality priority mode is a mode in which image quality is prioritized over other performances (rendering speed, etc.). The speed priority mode is a mode in which rendering speed is prioritized over other performances (image quality, etc.). A specific description will be given below.

[0030] [3. Example of System Configuration in Image Quality Priority Mode] In image quality priority mode, processing is performed to maximize other performance (rendering speed, etc.) within the scope that satisfies the required image quality. The required image quality refers to the standard of image quality required by user settings or design specifications. The processing phases include a level of detail setting phase and a rendering phase. The level of detail setting phase optimizes the level of detail setting table LT based on the image quality evaluation of the restored image RI. The rendering phase generates video for viewing using the optimized level of detail setting table LT.

[0031] 8 is a diagram showing an example of the system configuration of the information processing device 1 used in the rendering phase. The information processing device 1 has a database unit 11, a viewpoint setting unit 12, a distance calculation unit 13, a detail level table setting unit 14, a detail level setting unit 15, a rendering unit 16, and a detail restoration processing unit 17.

[0032] The database unit 11 holds 3D assets of various objects OB (information on three-dimensional shape, information on materials, information on optical properties such as reflectance and transmittance, etc.) as object data. For each object OB, multiple 3D models MD with different numbers of polygons are prepared. The number of polygons determines the accuracy of the detailed structure. The degree of accuracy of the detailed structure is defined as the level of detail LV. The database unit 11 stores multiple 3D models MD with different levels of detail LV for each object OB.

[0033] The level of detail setting unit 15 selects a 3D model MD having an appropriate level of detail LV (appropriate level of detail) from the perspective of the required image quality as an appropriate model, and outputs the selected model to the rendering unit 16. The determination of appropriateness is made taking into account a balance with other performance (e.g., rendering speed). For example, the level of detail setting unit 15 acquires, as an appropriate model, a 3D model MD in which the detailed structure has been reduced to the maximum extent possible within the range in which the restored image RI satisfies the required image quality.

[0034] The criteria for determining appropriateness are affected by factors such as the arrangement of the object OB. For example, for distant objects or small objects, lowering the level of detail LV does not make the roughness of the shape more noticeable. Conversely, for close objects or large objects, lowering the level of detail LV makes the roughness of the shape more noticeable. Therefore, the level of detail table setting unit 14 sets the level of detail LV required for the restored image RI to satisfy the required image quality as the appropriate level of detail for each item that affects the required image quality of the object OB.

[0035] For example, the level of detail table setting unit 14 includes a level of detail setting table LT. The level of detail setting table LT can include the distance DS from the rendering viewpoint PV to the object OB as an item that affects the required image quality. The level of detail setting table LT stores the distance DS and an appropriate level of detail for each object OB in association with each other. The level of detail setting unit 15 refers to the level of detail setting table LT and acquires from the database unit 11 a 3D model MD (appropriate model) having an appropriate level of detail corresponding to the distance DS.

[0036] The rendering viewpoint PV and the distance DS are acquired by the viewpoint setting unit 12 and the distance calculation unit 13. For example, the viewpoint setting unit 12 sets the rendering viewpoint PV of the object OB for each frame based on scene information that defines the camera angle, etc. The distance calculation unit 13 calculates the distance DS from the rendering viewpoint PV to the object OB for each frame based on the arrangement of the object OB in the rendering scene.

[0037] The rendering unit 16 performs rendering processing on the appropriate model acquired from the level of detail setting unit 15. The level of detail LV of the appropriate model is set so that the image after the detail restoration processing (restored image RI) satisfies the required image quality. Therefore, the appropriate model is a high-detail 3D model MD having the highest level of detail LV. Ｈ 3D model MD with a lower level of detail than (low-detail 3D model MD Ｌ The rendering unit 16 generates a low-detail 3D model MD with reduced detailed structures. Ｌ Rendering is performed on the low-detail 2D image IM Ｌ Get.

[0038] The rendering process can include viewpoint conversion processing based on the rendering viewpoint PV and ray simulation processing. The viewpoint conversion processing includes coordinate conversion processing according to camera work. The ray simulation processing includes processing (ray tracing) that simulates the propagation of light rays based on the laws of physics and reproduces the appearance of an object. Since the ray simulation processing requires complex three-dimensional calculations, it is difficult to create a highly detailed 3D model MD. Ｈ In this disclosure, the calculation target is a low-detail 3D model MD. Ｌ Therefore, the amount of calculation is significantly reduced.

[0039] The detail restoration processing unit 17 processes the low-detail 2D image IM obtained by the rendering process. Ｌ The detail restoration process uses an AI learning-based inference model such as GAN or Transformer to generate a 2D image IM (low-detail 2D image IM) of an object OB whose detailed shape is simplified. Ｌ ) to generate a 2D image IM of the object OB in which the detailed shape is accurately reproduced (high-detail 2D image IM Ｈ The detail restoration processing unit 17 restores the low-detail 2D image IM Ｌ From the image, a restored image RI (high-detail 2D image IM) in which the detailed structure is restored is obtained. Ｈ ) to generate the

[0040] 9 is a diagram showing an example of the system configuration of the information processing device 1 used in the level of detail setting phase. In addition to the configuration of FIG. 8, the information processing device 1 has a standard level of detail setting unit 21, an image quality evaluation index value calculation unit 22, and a table update unit 23.

[0041] In the detail level setting phase, the appropriate detail level of the detail level setting table LT is set based on the image quality evaluation of the restored image RI. Ｃ The standard level of detail setting unit 21 sets the level of detail LV, which is the standard for image quality evaluation, as the standard level of detail. The standard level of detail can be set arbitrarily by the system developer.

[0042] The detail level setting unit 15 selects a 3D model MD (reference model MD) having a reference detail level from the database unit 11. Ｃ ) and outputs it to the rendering unit 16. The image quality evaluation index value calculation unit 22 extracts the reference model MD Ｃ The rendering image of Ｃ The image quality evaluation index value calculation unit 22 acquires the reference image IM Ｃ and the restored image RI to calculate an image quality evaluation index value IV.

[0043] As the image quality evaluation index value IV, known indices such as PSNR (Peak signal-to-noise ratio), SSIM (Structural Similarity), and LPISP (Learned Perceptual Image Patch Similarity) can be adopted. For example, the image quality evaluation index value calculation unit 22 selectively performs image evaluation on the image range in which the object OB is displayed.

[0044] The image quality evaluation index value calculation unit 22 calculates the reference image IM Ｃ The image quality evaluation index value IV indicating the comparison result between the image quality evaluation index value IV and the restored image RI is obtained as an index for setting the appropriate level of detail. The table update unit 23 registers the setting of the level of detail LV optimized based on the image quality evaluation index value IV in the level of detail setting table LT for setting the appropriate level of detail.

[0045] Reference image IMＣ is the standard for image quality evaluation, so the reference image IM Ｃ It is desirable that the reference level of detail be an image that accurately reproduces the object OB. For example, the reference level of detail setting unit 21 sets the highest level of detail of multiple levels of detail set for the same object OB as the reference level of detail for the object OB. In the example of FIG. 2, this corresponds to "Lv0."

[0046] The table update unit 23 updates the reference image IM Ｃ The image quality of the image is acquired as the required image quality at a level equivalent to the image quality of the reference image IMC. The table update unit 23 sets the level of detail LV at which the required image quality is obtained as the appropriate level of detail. "Equal level" means a level that can be considered roughly equivalent to the image quality of the reference image IMC. The range of image quality that indicates the "equal level" can be set using a threshold value or the like. The range of the "equal level" can be set as desired by the system developer.

[0047] In the image quality priority mode, the level of detail optimized based on the required image quality is registered as the appropriate level of detail. For example, the table update unit 23 registers the setting that results in the lowest level of detail within the range that satisfies the required image quality in the image quality priority mode as the optimized level of detail. This minimizes the amount of rendering calculation within the range that satisfies the required image quality.

[0048] 10 and 11 are diagrams showing an example of the processing flow of the level of detail setting phase.

[0049] In the detail level setting phase, appropriate levels of detail are set for all pre-set objects OB and all camerawork (rendering viewpoints PV). First, the detail level setting unit 15 selects an arbitrary object OB registered in the database unit 11 as a rendering target (step S1). The viewpoint setting unit 12 selects an arbitrary rendering viewpoint PV (step S2).

[0050] The standard level of detail setting unit 21 sets the level of detail LV (standard level of detail) that serves as a standard for image quality to "Lv0" (step S3). The standard level of detail setting unit 21 retrieves the 3D model MD of the object OB corresponding to the standard level of detail from the database unit 11 as the reference model MD. Ｃand outputs it to the rendering unit 16. The rendering unit 16 obtains the reference model MD Ｃ Rendering is performed on the reference image IM Ｃ is generated (step S4).

[0051] The distance calculation unit 13 calculates the distance DS from the rendering viewpoint PV to the object OB based on the arrangement of the object OB in the rendering scene (step S5). Ｃ All 3D model MDs other than the above are 3D model MDs that are subject to the appropriate level of detail setting (low-detail 3D model MDs Ｌ ) to identify it.

[0052] The detail level setting unit 15 reads out the detail level setting table LT in which the default detail level setting has been performed. For example, the detail level setting table LT has the setting value of the detail level LV used in the conventional system of FIG. 3 registered as a default value. The detail level setting unit 15 refers to the detail level setting table LT and selects an arbitrary low-detail 3D model MD from the database unit 11. Ｌ is acquired as a rendering target and output to the rendering unit 16 (step S6).

[0053] The rendering unit 16 renders the acquired low-detail 3D model MD Ｌ Rendering is performed on the low-detail 2D image IM Ｌ (Step S7). The detail restoration processing unit 17 generates a low-detail 2D image IM Ｌ The image quality evaluation index value calculation unit 22 applies a detail restoration process to the image RI to generate a restored image RI in which the detailed structure is restored (step S8). Ｃ and calculates the image quality evaluation index value IV (step S9).

[0054] The image quality evaluation index value calculation unit 22 determines whether the image quality evaluation index value IV is equal to or less than a threshold value (step S10). Ｃ The threshold value can be set arbitrarily by the system developer. If the image quality evaluation index value IV is equal to or less than the threshold value, the restored image RI is determined to satisfy the required image quality.

[0055] If the image quality evaluation index value IV is greater than the threshold value (step S10: No), the process returns to step S7. The above-described process is then repeated until the image quality evaluation index value IV becomes equal to or less than the threshold value. If the image quality evaluation index value IV is equal to or less than the threshold value (step S10: Yes), the table update unit 23 changes the appropriate level of detail registered in the level of detail setting table LT to the next lower level of detail LV (step S11).

[0056] The viewpoint setting unit 12 determines whether the appropriate level of detail has been set for all rendering viewpoints PV for the selected object OB (step S12). If there is a rendering viewpoint PV for which the appropriate level of detail has not been set (step S12: No), the process returns to step S2. The above process is then repeated until the appropriate level of detail has been set for all rendering viewpoints PV.

[0057] If the setting of the appropriate level of detail has been completed for all rendering viewpoints PV (step S12: Yes), the level of detail setting unit 15 determines whether the setting of the appropriate level of detail has been completed for all objects OB (step S13). If there is an object OB for which the setting of the appropriate level of detail has not been completed (step S13: No), the process returns to step S1. Then, the above-mentioned process is repeated until the setting of the appropriate level of detail has been completed for all objects OB.

[0058] When the appropriate level of detail has been set for all objects OB (step S13: Yes), the level of detail setting phase is completed. Through the above process, the default level of detail setting is updated to a level of detail setting that takes into account the image quality evaluation results of the restored image RI. The updated level of detail setting is a level of detail setting that is appropriate for this technique involving detail restoration processing.

[0059] 12 is a diagram showing an example of the level of detail settings before and after the update. In the default level of detail settings, the appropriate levels of detail are "Level 0," "Level 1," "Level 2," and "Level 3" when the distance DS is "less than 3," "3 or more but less than 10," "10 or more but less than 30," and "30 or more," respectively. The appropriate levels of detail after the update are "Level 1," "Level 3," "Level 4," and "Level 5," respectively, which are the same as or lower than the levels before the update.

[0060] [3-3. Effects] The information processing device 1 has a rendering unit 16 and a detail restoration processing unit 17. The rendering unit 16 generates a low-detail 3D model MD with reduced detailed structures. Ｌ The detail restoration processing unit 17 performs rendering processing on the low-detail 2D image IM obtained by the rendering processing. Ｌ A restored image RI in which the detailed structure is restored is generated from the image data. In the information processing method disclosed herein, the processing of the information processing device 1 is executed by a computer. A computer-readable non-transitory storage medium disclosed herein stores a program that causes a computer to realize the processing of the information processing device 1.

[0061] According to this configuration, rendering is performed on a coarse 3D model MD with reduced detailed structure. This significantly reduces the amount of rendering calculations. Although processing to restore the detailed shape (detail restoration processing) is required after rendering, the detail restoration processing is performed on the 2D image IM, so the amount of calculations is less than that of the calculations for the 3D model MD. This method can also be applied to foreground objects OB. This reduces the amount of rendering calculations in a variety of scenes.

[0062] The rendering process includes a viewpoint conversion process based on the rendering viewpoint PV and a light simulation process.

[0063] This configuration reduces the processing load, particularly for ray simulations that require a large amount of calculation.

[0064] The information processing device 1 has a database unit 11 and a level of detail table setting unit 14. The database unit 11 stores, for each object OB, a plurality of 3D models MD with different levels of detail LV. The level of detail table setting unit 14 sets, as an appropriate level of detail, a level of detail LV necessary for the restored image RI to satisfy the required image quality for each item that affects the required image quality of the object OB.

[0065] This configuration makes it possible to reduce the amount of rendering calculation while maintaining the required image quality.

[0066] The level of detail table setting unit 14 includes a level of detail setting table LT that defines an appropriate level of detail. The level of detail setting table LT includes the distance DS from the rendering viewpoint PV to the object OB as an item that affects the required image quality.

[0067] According to this configuration, an appropriate level of detail LV can be set according to the distance DS.

[0068] The information processing device 1 includes a reference level of detail setting unit 21 and an image quality evaluation index value calculation unit 22. The reference level of detail setting unit 21 sets a level of detail LV, which is a standard for image quality evaluation, as a reference level of detail. The image quality evaluation index value calculation unit 22 calculates a rendering image of a 3D model MD having the reference level of detail as a reference image IM. Ｃ The image quality evaluation index value calculation unit 22 acquires the reference image IM Ｃ and the restored image RI, an image quality evaluation index value IV indicating the comparison result between the restored image RI and the restored image RI is acquired as a setting index for the appropriate level of detail.

[0069] According to this configuration, the reference image IM Ｃ An appropriate level of detail is set based on the image quality of the image.

[0070] The information processing device 1 has a table update unit 23. The table update unit 23 registers the setting of the level of detail LV optimized based on the image quality evaluation index value IV in a level of detail setting table LT for setting an appropriate level of detail.

[0071] According to this configuration, an appropriate level of detail LV is set according to the image quality of the restored image RI.

[0072] The reference level of detail setting unit 21 sets the highest level of detail LV of the plurality of levels of detail set for the same object OB as the reference level of detail of the object OB. Ｃ The table update unit 23 sets the level of detail LV at which the required image quality is obtained as the appropriate level of detail.

[0073] According to this configuration, a high-quality restored image RI can be obtained.

[0074] The table update unit 23 registers the setting that provides the lowest level of detail within the range that satisfies the required image quality of the image quality priority mode as the setting of the optimized level of detail.

[0075] According to this configuration, the amount of rendering calculations can be minimized while still satisfying the required image quality.

[0076] The effects described in this specification are merely examples and are not limiting, and other effects may also be present.

[0077] 4. Example of System Configuration in Speed ​​Priority Mode In speed priority mode, processing is performed to maximize other performance (image quality, etc.) within the scope of the required speed. The required speed refers to the standard rendering speed required by user settings or design specifications. The processing phases include an image quality evaluation index value measurement phase and a rendering phase. The image quality evaluation index value measurement phase generates an image quality evaluation index value table TB (see FIG. 14) based on the image quality evaluation of the restored image RI. The rendering phase generates video for viewing using the image quality evaluation index value table TB.

[0078] 13 is a diagram showing an example of the system configuration of an information processing device 2 used in the rendering phase. The information processing device 2 has a database unit 11, a viewpoint setting unit 12, a distance calculation unit 13, a level of detail setting unit 15, a rendering unit 16, a detail restoration processing unit 17, a rendering speed measurement unit 26, and an image quality evaluation index value table data extraction unit 27. Below, differences from the information processing device 1 shown in FIG. 8 will be described.

[0079] The rendering speed measurement unit 26 measures the number of frames that can be rendered per second (the number of restored images RI) as the rendering speed SP. The rendering speed SP is measured for the entire process, including both the rendering process and the detailed restoration process. The measurement means can be a means adopted in a general rendering engine.

[0080] The level of detail setting unit 15 acquires the level of detail LV that allows rendering within a range that satisfies the required speed (e.g., 30 FPS) as the appropriate level of detail. The rendering unit 16 acquires the 3D model MD having the appropriate level of detail from the database unit 11 and performs rendering processing.

[0081] The appropriate level of detail is determined by referring to partial table data TD extracted from the image quality evaluation index value table TB. Fig. 14 is a diagram showing an example of the image quality evaluation index value table TB. The image quality evaluation index value table TB records, for each object OB, an image quality evaluation index value IV for all items that affect the required image quality and all combinations of levels of detail LV. In the example of Fig. 14, the distance DS from the rendering viewpoint PV to the object OB is shown as an item that affects the required image quality. The image quality evaluation index value table data extraction unit 27 includes the image quality evaluation index value table TB as a data extraction target.

[0082] The image quality evaluation index value table data extraction unit 27 extracts data on some of the image quality evaluation index values ​​IV related to the rendering scene from the image quality evaluation index value table TB as partial table data TD. The level of detail setting unit 15 obtains, from the extracted partial table data TD, the level of detail LV of each object OB that provides the highest image quality within the range of the rendering speed SP (required speed) required in the speed-priority mode, as the appropriate level of detail for each object OV.

[0083] For example, the image quality evaluation index value table data extraction unit 27 acquires the distance DS of each object OB included in the rendering scene from the distance calculation unit 13. The image quality evaluation index value table data extraction unit 27 extracts, from the image quality evaluation index value table TB, data on the image quality evaluation index values ​​IV for all levels of detail LV corresponding to the extracted distance DS for each object OB, as partial table data TD.

[0084] FIG. 15 is a diagram showing an example of partial table data TD.

[0085] Fig. 15 shows an example in which data on some of the image quality evaluation index values ​​IV of "Object A" and "Object B" shown in Fig. 14 is extracted as partial table data TD. In the target rendering scene, "Object A" is placed at a distance DS of less than 3 from the rendering viewpoint PV. "Object B" is placed at a distance DS of 3 or more and less than 10 from the rendering viewpoint PV.

[0086] The image quality evaluation index value table data extraction unit 27 extracts image quality evaluation index value IV data for all "levels of detail" registered for "object name: Object A" and "distance: less than 3". The image quality evaluation index value table data extraction unit 27 extracts image quality evaluation index value IV data for all "levels of detail" registered for "object name: Object B" and "distance: 3 or more and less than 10". The level of detail setting unit 15 sorts the extracted data for each object OB in descending order of image quality evaluation index value IV.

[0087] When rendering a scene including "ObjectA" and "ObjectB", the level of detail setting unit 15 first sets the highest level of detail "Lv0" for both "ObjectA" and "ObjectB".

[0088] The rendering speed measurement unit 26 measures the rendering speed SP in the process of generating a restored image RI including "ObjectA" and "ObjectB." If the rendering speed SP meets the required speed, the level of detail setting unit 15 maintains the level of detail LV of "ObjectA" and "ObjectB" at "Lv0." If the rendering speed SP does not meet the required speed, the level of detail setting unit 15 lowers the level of detail LV of "ObjectA" or "ObjectB" by one level while minimizing degradation in image quality of the entire image.

[0089] The level of detail LV can be changed according to the arrangement order of the sorted partial table data TD. The sorted partial table data TD defines combinations of objects OB and level of detail LV in descending order of image quality evaluation index value IV. Therefore, by selecting the level of detail LV configurations in descending order from the top row of the partial table data TD, it is possible to appropriately change the level of detail LV of each object OB while minimizing degradation in the image quality of the entire image.

[0090] For example, referring to the sorted partial table data TD in Fig. 15, the first row describes a configuration for lowering the level of detail of "Object A" by one level. Therefore, the level of detail setting unit 15 changes the level of detail of "Object A" to "Level 1," which is one level lower, and maintains the level of detail of "Object B" at "Level 0."

[0091] The rendering speed measurement unit 26 measures the rendering speed SP in the generation process of the restored image RI after the setting of the level of detail LV has been changed. If the rendering speed SP satisfies the required speed, the level of detail setting unit 15 maintains the setting of the level of detail LV of each object OB. If the rendering speed SP does not satisfy the required speed, the level of detail setting unit 15 refers to the configuration of the second stage of the partial table data TD, and changes the level of detail LV of "Object B" to "Lv1", one level lower, while maintaining the level of detail LV of "Object A" at "Lv1".

[0092] The information processing device 2 continues the same process until the rendering speed SP meets the required speed. While the setting of the level of detail LV is being changed based on the partial table data TD, the rendering speed SP does not meet the required speed (e.g., 30 FPS), and therefore the user may be presented with an unsmooth image. However, if the setting of the level of detail LV can be optimized after a few changes, the time during which the user feels uncomfortable is likely to be minimized.

[0093] FIG. 16 is a diagram illustrating an example of a processing flow of the rendering phase.

[0094] The viewpoint setting unit 12 sets a rendering viewpoint PV based on scene information (step S21). The scene information specifies the placement of objects OB in the rendering scene, the camera angle, etc. The rendering unit 16 selects one object OB included in the rendering scene as a rendering target (step S22). The distance calculation unit 13 calculates a distance DS from the rendering viewpoint PV to the object OB based on the placement of the object OB in the rendering scene (step S23).

[0095] The image quality evaluation index value table data extraction unit 27 extracts image quality evaluation index value IV data for each distance DS of the selected object OB from the image quality evaluation index value table TB (step S24). The image quality evaluation index value table data extraction unit 27 determines whether or not there are other objects OB included in the rendering scene (step S25). If there are other objects OB included in the rendering scene (step S25: No), the process returns to step S22, and the above-mentioned process is repeated until image quality evaluation index value IV data for all objects OB has been extracted.

[0096] When the image quality evaluation index value IV data for all objects OB has been extracted (step S25: Yes), the image quality evaluation index value table data extraction unit 27 sorts the data extracted from the image quality evaluation index value table TB in descending order of image quality evaluation index value IV (step S26). In this way, the image quality evaluation index value table data extraction unit 27 acquires partial table data TD that defines combinations of objects OB and levels of detail LV in descending order of image quality evaluation index value IV.

[0097] The level of detail setting unit 15 sets the level of detail LV of all objects OB to the initial value "Lv0" (step S27). The rendering unit 16 applies rendering processing to the 3D model MD of each object OB having the set level of detail LV, and generates a 2D image IM of each object OB (step S28).

[0098] The detail restoration processing unit 17 generates a restored image RI from the 2D image IM, which is the input image (step S29). Ｌ However, if the input image is a high-detail 2D image, the detail restoration processing is not performed and the input image is output as is as a restored image RI. Since the 3D model MD with the highest level of detail LV, "Lv0", is the target for rendering the first time, the detail restoration processing unit 17 does not apply the detail restoration processing to the 2D image IM and outputs it as is as a restored image RI.

[0099] The rendering speed measurement unit 26 measures the rendering speed SP based on the time from the start of the rendering process to the output of the restored image RI (step S30). The rendering speed measurement unit 26 compares the rendering speed SP with a threshold (required speed) (step S31). If the rendering speed SP is equal to or greater than the threshold (step S31: Yes), the level of detail setting unit 15 maintains the level of detail LV of each object OB and completes the level of detail LV setting process.

[0100] If the rendering speed SP is less than the threshold value (step S31: No), the level of detail setting unit 15 refers to the partial table data TD and selects a combination of levels of detail LV for each object OB in descending order of image quality evaluation index value IV (step S32).Then, the process returns to step S28, and the above-described process is repeated until the rendering speed SP becomes equal to or greater than the threshold value.

[0101] 17 is a diagram showing an example of the system configuration of an information processing device 2 used in the image quality evaluation index value measurement phase. In addition to the configuration of FIG. 13, the information processing device 2 has a reference level of detail setting unit 21, an image quality evaluation index value calculation unit 22, and an image quality evaluation index value table creation unit 25. The reference level of detail setting unit 21 and the image quality evaluation index value calculation unit 22 are the same as those shown in FIG. 9.

[0102] In the image quality evaluation index value measurement phase, an image quality evaluation index value table TB is created based on the image quality evaluation of the restored image RI. Ｃ This is done based on a comparison with the image quality of the reference image IM Ｃ is a 3D model MD (reference model MD) having a level of detail LV (reference level of detail) set by the reference level of detail setting unit 21. Ｃ The image quality evaluation index value calculation unit 22 calculates the image quality evaluation index value by rendering the reference image IM Ｃ and the restored image RI to calculate an image quality evaluation index value IV.

[0103] The image quality evaluation index value table creation unit 25 records, for each object OB, an image quality evaluation index value IV for all combinations of all items that affect the required image quality and all levels of detail LV. The image quality evaluation index value table creation unit 25 obtains the recording results as an image quality evaluation index value table TB. In the example of Figure 17, the distance DS from the rendering viewpoint PV to the object OB is shown as an item that affects the required image quality.

[0104] 18 and 19 are diagrams showing an example of the processing flow of the image quality evaluation index value measurement phase.

[0105] In the image quality evaluation index value measurement phase, the image quality of the restored image RI is evaluated for all combinations of objects OB, all cameraworks (rendering viewpoints PV), and all levels of detail LV. First, the level of detail setting unit 15 selects an arbitrary object OB registered in the database unit 11 as a rendering target (step S41). The viewpoint setting unit 12 selects an arbitrary rendering viewpoint PV (step S42).

[0106] The standard level of detail setting unit 21 sets the level of detail LV (standard level of detail) that is the standard for image quality to "Lv0" (step S43). The standard level of detail setting unit 21 retrieves the 3D model MD of the object OB that corresponds to the standard level of detail from the database unit 11 as the reference model MD. Ｃ and outputs it to the rendering unit 16. The rendering unit 16 obtains the reference model MD Ｃ Rendering is performed on the reference image IM Ｃ is generated (step S44).

[0107] The distance calculation unit 13 calculates the distance DS from the rendering viewpoint PV to the object OB based on the arrangement of the object OB in the rendering scene (step S45). Ｃ All 3D model MDs other than the above are subject to image quality evaluation. Ｌ ) to identify it.

[0108] The detail level setting unit 15 sets the detail level LV of the 3D model MD to be rendered to "Lv1" (step S46). Ｌ from the database unit 11 and outputs it to the rendering unit 16. The rendering unit 16 converts the acquired low-detail 3D model MD Ｌ Rendering is performed on the low-detail 2D image IM Ｌ is generated (step S47).

[0109] The detail restoration processing unit 17 generates a low-detail 2D image IM ＬThe image quality evaluation index value calculation unit 22 applies a detail restoration process to the image RI to generate a restored image RI in which the detailed structure is restored (step S48). Ｃ The image quality evaluation index value table creating section 25 compares the image quality evaluation index value IV with the image quality evaluation index value IV and records the calculated image quality evaluation index value IV in the image quality evaluation index value table TB (step S50).

[0110] The detail level setting unit 15 determines whether image quality evaluation has been completed for all detail levels for the selected object OB and rendering viewpoint PV (step S51). If there is a detail level for which image quality evaluation has not been completed (step S51: No), the process returns to step S47. The above-described process is then repeated until image quality evaluation has been completed for all detail levels.

[0111] If image quality evaluation has been completed for all levels of detail LV (step S51: Yes), the viewpoint setting unit 12 determines whether image quality evaluation has been completed for all rendering viewpoints PV for the selected object OB (step S52). If there is a rendering viewpoint PV for which image quality evaluation has not been completed (step S52: No), the process returns to step S42. The above-described process is then repeated until image quality evaluation has been completed for all rendering viewpoints PV.

[0112] If the image quality evaluation has been completed for all rendering viewpoints PV (step S52: Yes), the level of detail setting unit 15 determines whether the image quality evaluation has been completed for all objects OB (step S53). If there are objects OB for which the image quality evaluation has not been completed (step S53: No), the process returns to step S41. The above-described process is then repeated until the image quality evaluation has been completed for all objects OB.

[0113] When image quality evaluation has been completed for all objects OB (step S53: Yes), the image quality evaluation index value measurement phase is completed. Through the above processing, an image quality evaluation index value table TB such as that shown in FIG.

[0114] [4-3. Effects] In the above-described processing, the level of detail of each object OB is appropriately set based on the image quality evaluation index value table TB. The image quality evaluation index value table data extraction unit 27 extracts data on some of the image quality evaluation index values ​​IV related to the rendering scene from the image quality evaluation index value table TB as partial table data TD. The level of detail setting unit 15 obtains, from the extracted partial table data TD, the level of detail LV of each object OB that provides the highest image quality within the range of the rendering speed SP determined in speed-priority mode, as the appropriate level of detail for each object OB.

[0115] According to this configuration, the highest possible image quality can be obtained within the range in which the necessary rendering speed SP can be ensured.

[0116] The image quality evaluation index value table data extraction unit 27 sorts the data extracted from the image quality evaluation index value table TB and acquires partial table data TD that defines combinations of objects OB and levels of detail LV in descending order of image quality evaluation index value IV. The level of detail setting unit 15 refers to the partial table data TD and selects combinations of levels of detail LV for each object OB in descending order of image quality evaluation index value IV.

[0117] According to this configuration, the combination of appropriate levels of detail can be determined automatically.

[0118] 5. Example of Hardware Configuration FIG. 20 is a diagram showing an example of the hardware configuration of the information processing devices 1 and 2. As shown in FIG.

[0119] The information processing of the information processing devices 1 and 2 is realized by, for example, a computer 1000. The computer 1000 has a CPU (Central Processing Unit) 1100, a RAM (Random Access Memory) 1200, a ROM (Read Only Memory) 1300, a HDD (Hard Disk Drive) 1400, a communication interface 1500, and an input / output interface 1600. The components of the computer 1000 are connected by a bus 1050.

[0120] The CPU 1100 operates and controls each component based on a program (program data 1450) stored in the ROM 1300 or the HDD 1400. For example, the CPU 1100 loads the program stored in the ROM 1300 or the HDD 1400 into the RAM 1200 and executes processing corresponding to the various programs.

[0121] The ROM 1300 stores boot programs such as a Basic Input Output System (BIOS) that is executed by the CPU 1100 when the computer 1000 starts up, as well as programs that depend on the hardware of the computer 1000 .

[0122] The HDD 1400 is a non-transitory computer-readable recording medium that non-temporarily records programs executed by the CPU 1100 and data used by such programs. Specifically, the HDD 1400 is a recording medium that records an information processing program according to an embodiment as an example of program data 1450.

[0123] The communication interface 1500 is an interface for connecting the computer 1000 to an external network 1550 (e.g., the Internet). For example, the CPU 1100 receives data from other devices and transmits data generated by the CPU 1100 to other devices via the communication interface 1500.

[0124] The input / output interface 1600 is an interface for connecting the input / output device 1650 and the computer 1000. For example, the CPU 1100 receives data from an input device such as a keyboard or a mouse via the input / output interface 1600. The CPU 1100 also transmits data to an output device such as a display device, a speaker, or a printer via the input / output interface 1600. The input / output interface 1600 may also function as a media interface for reading programs recorded on a predetermined recording medium. Examples of media include optical recording media such as a DVD (Digital Versatile Disc) or a PD (Phase Change Rewritable Disc), magneto-optical recording media such as an MO (Magneto-Optical Disk), tape media, magnetic recording media, and semiconductor memories.

[0125] For example, when the computer 1000 functions as the information processing devices 1 and 2 according to the embodiments, the CPU 1100 of the computer 1000 executes an information processing program loaded onto the RAM 1200 to realize the functions of the aforementioned components. The information processing program, various models, and various data according to the present disclosure are stored in the HDD 1400. The CPU 1100 reads and executes program data 1450 from the HDD 1400. Alternatively, the CPU 1100 may acquire these programs from another device via an external network 1550.

[0126] [Additional Notes] The present technology may also have the following configurations. (1) An information processing device having: a rendering unit that performs rendering processing on a low-detail 3D model in which a detailed structure has been reduced; and a detail restoration processing unit that generates a restored image in which the detailed structure has been restored from a low-detail 2D image obtained by the rendering processing. (2) The information processing device according to (1), wherein the rendering processing includes viewpoint conversion processing based on a rendering viewpoint and ray simulation processing. (3) The information processing device according to (1) or (2), wherein the information processing device has: a database unit that stores, for each object, a plurality of 3D models with different levels of detail; and a level-of-detail table setting unit that sets, for each item that affects the required image quality of the object, the level of detail necessary for the restored image to satisfy the required image quality as an appropriate level of detail. (4) The information processing device according to (3), wherein the level-of-detail table setting unit includes a level-of-detail setting table that defines the appropriate level of detail, and the level-of-detail setting table includes a distance from a rendering viewpoint to the object as one of the items that affect the required image quality. (5) The information processing device according to (3) or (4), comprising: a standard level of detail setting unit that sets a level of detail that serves as a standard for image quality evaluation as a standard level of detail; and an image quality evaluation index value calculation unit that acquires a rendering image of a 3D model having the standard level of detail as a standard image, and acquires an image quality evaluation index value that indicates a comparison result between the standard image and the restored image as an index for setting the appropriate level of detail. (6) The information processing device according to (5), comprising a table update unit that registers the level of detail setting optimized based on the image quality evaluation index value in a level of detail setting table for setting the appropriate level of detail. (7) The information processing device according to (6), wherein the standard level of detail setting unit sets the highest level of detail of multiple levels of detail set for the same object as the standard level of detail of the object, and the table update unit acquires an image quality at a level equivalent to the image quality of the standard image as the required image quality, and sets the level of detail that achieves the required image quality as the appropriate level of detail.(8) The information processing device according to (7), wherein the table update unit registers, as the optimized level of detail setting, a setting which provides the lowest level of detail within a range which satisfies the required image quality in image quality priority mode. (9) The information processing device according to (5), comprising: an image quality evaluation index value table data extraction unit which extracts, as partial table data, data of a part of the image quality evaluation index values ​​related to a rendering scene from an image quality evaluation index value table which records, for each of the objects, the image quality evaluation index values ​​for all items which affect the required image quality and all combinations of the levels of detail; and a level of detail setting unit which obtains, from the extracted partial table data, the level of detail of each object which provides the highest image quality within a range of the rendering speed required in speed priority mode, as the appropriate level of detail of each object. (10) The information processing device according to (9), wherein the image quality evaluation index value table data extraction unit sorts data extracted from the image quality evaluation index value table to obtain the partial table data defining combinations of the objects and the levels of detail in descending order of the image quality evaluation index value, and the level of detail setting unit refers to the partial table data to select the combinations of the levels of detail for each object in descending order of the image quality evaluation index value. (11) An information processing method executed by a computer, comprising: performing a rendering process on a low-detail 3D model from which a detailed structure has been reduced; and generating a restored image in which the detailed structure has been restored from a low-detail 2D image obtained by the rendering process. (12) A computer-readable non-transitory storage medium storing a program for causing a computer to perform the steps of: performing a rendering process on a low-detail 3D model from which a detailed structure has been reduced; and generating a restored image in which the detailed structure has been restored from a low-detail 2D image obtained by the rendering process.

[0127] 1, 2 Information processing device. 11 Database unit 14 Level of detail table setting unit 15 Level of detail setting unit 16 Rendering unit 17 Detail restoration processing unit 21 Reference level of detail setting unit 22 Image quality evaluation index value calculation unit 23 Table update unit 27 Image quality evaluation index value table data extraction unit DS Distance IM 2D image IM Ｃ Reference image IM Ｌ Low-detail 2D image IV Image quality evaluation index value LT Detail level setting table LV Detail level MD 3D model MD Ｌ Low-detail 3D model OB Object PV Rendering viewpoint RI Reconstructed image SP Rendering speed TB Image quality evaluation index value table TD Partial table data

Claims

1. An information processing device having: a rendering unit that performs a rendering process on a low-detail 3D model in which detailed structures have been reduced; and a detail restoration processing unit that generates a restored image in which the detailed structures have been restored from a low-detail 2D image obtained by the rendering process.

2. The information processing device according to claim 1, wherein the rendering process includes a viewpoint conversion process based on a rendering viewpoint and a light simulation process.

3. An information processing device as described in claim 1, comprising: a database unit that stores multiple 3D models with different levels of detail for each object; and a level of detail table setting unit that sets the level of detail required for the restored image to satisfy the required image quality as the appropriate level of detail for each item that affects the required image quality of the object.

4. The information processing device according to claim 3, wherein the level of detail table setting unit includes a level of detail setting table that specifies the appropriate level of detail, and the level of detail setting table includes the distance from a rendering viewpoint to the object as an item that affects the required image quality.

5. An information processing device as described in claim 3, comprising: a standard detail level setting unit that sets a level of detail that serves as a standard for image quality evaluation as a standard detail level; and an image quality evaluation index value calculation unit that acquires a rendering image of a 3D model having said standard detail level as a standard image, and acquires an image quality evaluation index value that indicates a comparison result between said standard image and said restored image as a setting index for said appropriate detail level.

6. The information processing device according to claim 5, further comprising a table update unit that registers the level of detail setting optimized based on the image quality evaluation index value in the level of detail setting table for setting the appropriate level of detail.

7. The information processing device described in claim 6, wherein the standard level of detail setting unit sets the highest level of detail among multiple levels of detail set for the same object as the standard level of detail for the object, and the table update unit obtains an image quality at a level equivalent to the image quality of the reference image as the required image quality, and sets the level of detail at which the required image quality is obtained as the appropriate level of detail.

8. The information processing device according to claim 7, wherein the table update unit registers, as the optimized setting of the level of detail, a setting that provides the lowest level of detail within a range that satisfies the required image quality in the image quality priority mode.

9. An information processing device as described in claim 5, comprising: an image quality evaluation index value table data extraction unit that extracts, as partial table data, data on some of the image quality evaluation index values ​​related to a rendering scene from an image quality evaluation index value table that records, for each of the objects, the image quality evaluation index values ​​for all items that affect the required image quality and all combinations of the levels of detail; and a level of detail setting unit that obtains, from the extracted partial table data, the level of detail of each object that provides the highest image quality within the range of the rendering speed required in speed priority mode, as the appropriate level of detail of each object.

10. The information processing device of claim 9, wherein the image quality evaluation index value table data extraction unit sorts the data extracted from the image quality evaluation index value table to obtain the partial table data that specifies combinations of the objects and the levels of detail in order of highest image quality evaluation index value, and the level of detail setting unit refers to the partial table data and selects the combinations of the levels of detail for each object in order of highest image quality evaluation index value.

11. An information processing method executed by a computer, comprising: performing a rendering process on a low-detail 3D model in which detailed structures have been reduced; and generating a restored image in which the detailed structures have been restored from a low-detail 2D image obtained by the rendering process.

12. A computer-readable non-transitory storage medium storing a program that causes a computer to perform the following steps: rendering a low-detail 3D model in which detailed structures have been reduced; and generating a restored image in which the detailed structures have been restored from a low-detail 2D image obtained by the rendering process.

Images