Multi-view video generation device and its program

Abstract

To provide a multi-view video generation apparatus capable of generating a multi-view video from a 3D model for CG more simply at a higher speed than conventional ones.SOLUTION: A multi-view video generation apparatus 1 includes: 3D model deformation means 12 which enlarges a 3D model in a direction parallel to an imaging surface of a virtual camera, in camera coordinate systems of multiple virtual cameras virtually arranged and separated by a predetermined visual range of a three-dimensional video display device from an enlargement start point corresponding to multiple viewpoint videos, to deform the 3D model so as to be imaged by a corresponding virtual camera; virtual camera imaging means 13 which virtually images the 3D model deformed by the 3D model deformation means 12 for each of the camera coordinate systems of the virtual cameras, using the virtual cameras, as viewpoint images; and multi-view video output means 14 which outputs the multiple viewpoint images imaged by the virtual camera imaging means 13, as a multi-view video.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a multi-view video generation device that generates multi-view videos from a CG 3D model and a program thereof.

Background Art

[0002] In recent years, various three-dimensional video display methods have been proposed, including binocular types using 3D glasses. In particular, a three-dimensional video display method that reproduces a high-density light beam group by superimposing and irradiating a multi-view video from the back onto a diffuser screen does not require special glasses and can display a natural three-dimensional video with horizontal and vertical parallax (see Patent Document 1).

[0003] However, in this method, since the light of each view video constituting the multi-view video is condensed at the focal position of the imaging lens and irradiated onto the diffuser screen, the view video cannot be defined by a general camera model such as perspective projection or orthographic projection. Therefore, it is difficult to generate a view video (multi-view video) that becomes a 3D CG scene from a CG 3D model by this method.

[0004] Among the methods for generating a video of a 3D scene from an elemental image group of the integral 3D method, there is a method (oblique projection method) of generating an elemental image by rearranging each pixel of an image group photographed by a virtual camera of oblique projection (see Patent Document 2, Non-Patent Document 1). In addition, as another method for generating a video of a 3D scene from an elemental image group of the integral 3D method, there is a method (ray tracing method) of generating an elemental image by obtaining each pixel of the elemental image one by one by ray tracing (see Patent Document 3). These methods for generating a video of a 3D scene from an elemental image group of the integral 3D method can be used as a method for generating a video of a 3D CG scene with a multi-view video.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

[0006] [Non-Patent Document 1] Y. Iwadate and M. Katayama: “Generating Integral Image from 3D Object by Using Oblique Projection,” Proc. IDW’11, ITE, Nagoya, 3Dp-1, pp.269-272(2011) [Summary of the Invention] [Problems to be Solved by the Invention]

[0007] As described above, in the three-dimensional video display method using multi-viewpoint video, it is difficult to generate a viewpoint video (multi-viewpoint video) that becomes a 3DCG scene from a 3D model of CG because a general camera model cannot be used.

[0008] Further, when the conventional oblique projection method is used for generating a multi-viewpoint video, it is necessary to shoot an image group with a virtual camera the same number of times as the number of pixels of the viewpoint video. For example, when the viewpoint video has 300,000 pixels, this method requires shooting 300,000 times from different directions as a 3DCG scene with a virtual camera. In addition, this method requires rearranging the pixels of the captured image to generate a viewpoint video. Therefore, this method has a problem that it takes a long time to generate a multi-viewpoint video.

[0009] Further, when the conventional ray tracing method is used for generating a multi-viewpoint video, a hit determination process between each object in the 3D model of CG and a ray is required, and the process becomes complicated. Therefore, this method has a problem that it takes a long time to generate a multi-viewpoint video when the number of polygons constituting the 3D model is large.

[0010] The present invention has been made in view of such problems, and it is an object of the present invention to provide a multi-view video generation device and a program thereof that can generate a viewpoint video by a single shooting with a virtual camera from a 3D model of CG and can generate a multi-view video more simply and at a higher speed than in the prior art.

Means for Solving the Problems

[0011] In order to solve the above problems, a multi-view video generation device according to the present invention is a multi-view video generation device that generates the multi-view video used for a three-dimensional video display device that displays a three-dimensional video by condensing the video light of a plurality of viewpoint videos of the multi-view video, expanding the video light starting from the condensing point, and superimposing the expanded video light on a diffusion screen, and is configured to include 3D model deformation means, virtual camera shooting means, and multi-view video output means.

[0012] In such a configuration, the multi-view video generation device, by the 3D model deformation means, in the camera coordinate system of each of a plurality of virtual cameras virtually arranged at a predetermined viewing distance from the expansion starting point corresponding to a plurality of viewpoint videos in a direction parallel to the imaging plane of the virtual camera, expands the 3D model and deforms the 3D model so as to be imaged by the corresponding virtual camera. In this way, the 3D model deformation means can deform the 3D model by a simple linear transformation. As a result, the deformed 3D model can be projected onto the virtual camera by a general camera model.

[0013] Then, the multi-view video generation device, by the virtual camera shooting means, virtually shoots the 3D model deformed by the 3D model deformation means for each camera coordinate system of the plurality of virtual cameras as a viewpoint video with each virtual camera. In this way, the viewpoint videos shot by the virtual cameras corresponding to the expansion starting points are videos with different viewpoint positions.

[0014] Then, the multi-view video generation device outputs, as multi-view video, a plurality of view videos captured by the virtual camera shooting means by the multi-view video output means. As a result, the multi-view video generation device can easily deform the 3D model, can generate view videos by a single projection process, and can generate view videos at high speed.

[0015] Note that the multi-view video generation device can be operated by a multi-view video generation program for causing a computer to function as each of the above-described means.

Effect of the Invention

[0016] According to the present invention, by deforming a CG 3D model according to the positional relationship with a virtual camera, individual view videos can be generated based on a general projection transformation camera model. As a result, the present invention can generate view videos by a single shooting with a virtual camera, and can generate multi-view videos more simply and at higher speed than before.

Brief Description of the Drawings

[0017]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Embodiments for Carrying Out the Invention

[0018] <Configuration of Multi-Viewpoint Video Generation Device> First, with reference to FIG. 1, the configuration of a multi-viewpoint video generation device 1 according to an embodiment of the present invention will be described. The multi-viewpoint video generation device 1 generates a multi-viewpoint video that becomes a 3DCG scene to be displayed on a three-dimensional video display device from a 3D model of CG (Computer Graphics). This multi-viewpoint video generation device 1 condenses the image light of a plurality of viewpoint videos of the multi-viewpoint video respectively, expands the image light with the condensing point as the expansion starting point, and superimposes it on a diffusion screen to generate a multi-viewpoint video used for a three-dimensional video display device that displays a three-dimensional video.

[0019] The 3D model is data representing objects such as people and objects in a three-dimensional CG space (3DCG space). This 3D model is represented by, for example, polygon data, point cloud (point group data), etc. The 3DCG scene is a scene obtained by photographing the 3D model in the 3DCG space from a certain viewpoint. The multi-viewpoint video is a video composed of a plurality of videos (viewpoint videos) with different viewpoint positions in the horizontal and vertical directions.

[0020] As shown in FIG. 1, the multi-viewpoint video generation device 1 includes a storage means 10, a virtual camera setting means 11, a 3D model deformation means 12, a virtual camera photographing means 13, and a multi-viewpoint video output means 14.

[0021] The storage means 10 stores in advance the configuration data (display device configuration data) of a three-dimensional video display device that displays a three-dimensional video using the generated multi-viewpoint video. This storage means 10 can be configured with a general storage medium such as a semiconductor memory. The display device configuration data are parameters for configuring the three-dimensional video display device.

[0022] Here, with reference to FIG. 2, the display device configuration data will be described by taking as an example a general three-dimensional video display device 2 (see, for example, Patent Document 1) that displays a three-dimensional video using a multi-viewpoint video. As shown in FIG. 2, the three-dimensional video display device 2 displays a multi-viewpoint video I composed of a plurality of viewpoint videos i with different viewpoint positions by a multi-viewpoint video display unit 20.

[0023] The video light of the multi-viewpoint video I displayed by the multi-viewpoint video display unit 20 is condensed by an imaging lens array 21 composed of imaging lenses Li facing the individual viewpoint videos i, and is focused on respective apertures (openings) Ai of an aperture array 22 arranged at a distance of the focal length f1 of the imaging lens Li, enlarged, and irradiated onto a diffusion screen 23. The center of this aperture Ai is the starting point where the video light of the viewpoint video expands. At this time, the principal rays of the video light of the individual viewpoint videos i are perpendicular to the diffusion screen 23. Also, the respective viewpoint videos i are partially overlapped and irradiated onto the diffusion screen 23. Note that this partial overlap of the viewpoint videos i is in a state where two or more viewpoint videos i are overlapped and irradiated in all regions except the peripheral portion of the diffusion screen 23.

[0024] Then, the three-dimensional video display device 2 slightly diffuses the incident discrete light beam group by the diffusion screen 23 to form a light beam group with a continuous luminance distribution. As a result, the observer O can visually recognize the three-dimensional video T as an optical image.

[0025] Note that in the multi-viewpoint video I, the viewpoint videos i are two-dimensionally arranged in the horizontal and vertical directions. Also, the imaging lenses Li and the apertures Ai are two-dimensionally arranged facing the viewpoint videos i.

[0026] For example, as shown in FIG. 3, the multi-view video I is a video in which videos (viewpoint videos i) at different viewpoint positions in the horizontal and vertical directions are arranged in a single image. FIG. 3 shows an example of the multi-view video I composed of viewpoint videos i with 5 viewpoints horizontally and 5 viewpoints vertically. In practice, the individual viewpoint videos i are inverted in the horizontal and vertical directions before and after the aperture array 22. Here, for simplicity of explanation, the multi-view video I is represented by non-inverted videos.

[0027] Thus, in a general three-dimensional video display device 2 that displays a three-dimensional video using a multi-view video, the display device configuration data includes the number of viewpoint videos i (horizontal number, vertical number), the display pitch (horizontal pitch, vertical pitch) p of the viewpoint videos i, the viewing angle φ, the distance (recommended viewing distance) L from the aperture array 22 (expansion start surface) to the observer O, and the like. Returning to FIG. 1, the description of the configuration of the multi-view video generation device 1 will be continued.

[0028] The virtual camera setting means 11 sequentially sets virtual cameras that virtually photograph the 3D model with reference to the display device configuration data stored in the storage means 10. The virtual camera setting means 11 sequentially sets the positions of the virtual cameras at (0, 0) to (Nx - 1, Ny - 1) based on the number of viewpoint videos i (horizontal number, vertical number) in the display device configuration data. Here, Nx is the horizontal number of viewpoint videos, and Ny is the vertical number of viewpoint videos. For example, the virtual camera setting means 11 sets the position of the virtual camera at (0, 0) as an initial value and sequentially sets the position of the virtual camera up to (Nx - 1, Ny - 1). The virtual camera setting means 11 outputs the initial position of the virtual camera to the 3D model deformation means 12 and the virtual camera photographing means 13.

[0029] In addition, when the virtual camera setting means 11 is notified by the virtual camera photographing means 13 that the photographing by the virtual camera has ended, the virtual camera setting means 11 resets the position of the virtual camera to the position of the virtual camera corresponding to another viewpoint video, and outputs the set position of the virtual camera to the 3D model deformation means 12 and the virtual camera photographing means 13.

[0030] The 3D model deformation means 12 deforms the 3D model in the camera coordinate system corresponding to the position of the virtual camera set by the virtual camera setting means 11. This 3D model deformation means 12 enlarges the 3D model in the camera coordinate system of each of the plurality of virtual cameras virtually arranged at a distance of a predetermined viewing distance (recommended viewing distance) L of the three-dimensional video display device 2 (FIG. 2) from the starting point of enlargement of the video light corresponding to the plurality of viewpoint videos, and deforms the 3D model so as to be photographed by the corresponding virtual camera by enlarging it in a direction parallel to the imaging plane of the virtual camera. Note that since the position of the starting point of enlargement is the point where the video light of the viewpoint video converges, the positions of the individual starting points of enlargement are separated by the display pitch p (FIG. 2) of the viewpoint video on the same plane.

[0031] Here, with reference to FIG. 4, the deformation of the 3D model in the 3D model deformation means 12 will be specifically described. FIG. 4 is a diagram modeling the positional relationship between the video light of the viewpoint video of the three-dimensional video display device 2 in FIG. 2, the 3D model, and the virtual camera. The starting point of enlargement Pi in FIG. 4 corresponds to the center of the aperture Ai in FIG. 2. Further, the starting point plane P in FIG. 4 corresponds to the aperture array 22 in FIG. 2. In FIG. 4, the illustration of the light ray from the viewpoint video i to the starting point of enlargement Pi is omitted. The x-axis, y-axis, and z-axis are the camera coordinate system of the virtual camera C in which each axis is orthogonal. The imaging center Ci of the virtual camera C exists at the origin coordinates (0, 0, 0), and the positive direction of the z-axis is the front direction of the virtual camera C.

[0032] In the positional relationship shown in FIG. 4, the starting point of enlargement Pi exists on the z-axis, and the viewpoint video light Lti expands from the starting point of enlargement Pi at the viewing angle φ and is projected onto the virtual camera C on the xy plane with a width dx in the x direction and a width dy (not shown) centered on the imaging center Ci. The 3D model deformation means 12 deforms the 3D model M according to the following equations (1) and (2) in order to match the positional relationship shown in FIG. 4.

[0033]

Number

[0034] However, x i , y i , z i are the coordinates of the constituent points that make up the 3D model M, and are the x, y, z coordinates in the camera coordinate system. For example, if the 3D model M is polygon data, x i , y i , z i are the x, y, z coordinates in the camera coordinate system of the i-th vertex of the polygon group. Also, if the 3D model M is a point cloud, x i , y i , z i are the x, y, z coordinates in the camera coordinate system of each point. However, z i is such that 0 < z i < L. Also, L is the distance from the starting point of expansion Pi to the imaging center Ci of the virtual camera C, and is the viewing distance (recommended viewing distance) shown in FIG. 2. x i ′, y i ′ are the x, y coordinates after deformation.

[0035] Note that when the position of a new virtual camera is set, the 3D model deformation means 12 translates the camera coordinate system in the x direction or the y direction so that the starting point of expansion Pi corresponding to the new virtual camera exists on the z-axis, and sets the imaging center of the new virtual camera to the origin coordinates (0, 0, 0) of the camera coordinate system. Thereby, the 3D model deformation means 12 can deform the 3D model M corresponding to the position of the virtual camera. Returning to FIG. 1, the description of the configuration of the multi-viewpoint video generation apparatus 1 will be continued.

[0036] The 3D model deformation means 12 outputs the deformed 3D model to the virtual camera imaging means 13. This 3D model deformation means 12 deforms the 3D model M corresponding to the position of the virtual camera every time the position of a new virtualized camera is set by the virtual camera setting means 11.

[0037] The virtual camera photographing means 13 virtually photographs the 3D model deformed by the 3D model deformation means 12 from the position of the virtual camera set by the virtual camera setting means 11. This virtual camera photographing means 13 generates a viewpoint video by photographing, on the photographing plane of the virtual camera, a range specified by the viewing angle of the display device configuration data stored in the storage means 10. For example, as shown in FIG. 4, when the viewing angle is φ and the viewing distance (recommended viewing distance) is L, the virtual camera photographing means 13 sets the photographing range to d x (= 2L·tan(φ / 2)) in the x direction with the photographing center Ci of the virtual camera C as the center. The same applies to the y direction.

[0038] Virtually photographing the 3D model by the virtual camera photographing means 13 means projecting the deformed 3D model onto the photographing plane of the virtual camera by performing an orthographic transformation. For example, when the 3D model M is polygon data, the virtual camera photographing means 13 may render each polygon of the deformed 3D model M one by one by orthographic transformation using a general 3D CG graphic interface such as OpenGL (registered trademark) or DirectX (registered trademark). Alternatively, when the virtual camera photographing means 13 uses a game engine such as Unity (registered trademark), the deformed 3D model M may be rendered by orthographic transformation using the vertex shader function.

[0039] Also, when the 3D model M is a point cloud (point group data), the virtual camera photographing means 13 may project each point of the deformed 3D model M onto the photographing plane of the virtual camera by orthographic transformation and draw each point with an arbitrary size.

[0040] The virtual camera photographing means 13 outputs the viewpoint video generated by virtual photographing with the virtual camera to the multi-viewpoint video output means 14. After generating one viewpoint video, the virtual camera photographing means 13 notifies the virtual camera setting means 11 that the photographing by the virtual camera has ended.

[0041] The multi-view video output means 14 outputs a plurality of view videos generated by virtual shooting by the virtual camera shooting means 13 as a multi-view video. The multi-view video output means 14 generates a single multi-view video by arranging the view videos in the horizontal and vertical directions according to the number of view videos (horizontal number, vertical number) stored in the storage means 10. Of course, the multi-view video output means 14 does not necessarily have to generate a multi-view video as a single video, and may output the view videos in a predetermined order in sequence as a multi-view video.

[0042] As described above, the multi-view video generation device 1 can convert a 3D model into a shape that can be orthogonally projected according to the position of the virtual camera. As a result, the multi-view video generation device 1 can generate a view video in one shot according to the position of the virtual camera, and can generate a multi-view video more simply and at a higher speed than before. Note that the multi-view video generation device 1 can be operated by a program (multi-view video generation program) that causes a computer to function as each of the above-described means.

[0043] <Operation of the multi-view video generation device> Next, with reference to FIG. 5 (refer to FIG. 1 as appropriate for the configuration), the operation of the multi-view video generation device 1 according to the embodiment of the present invention will be described. It is assumed that the storage means 10 stores in advance configuration data (display device configuration data) of a three-dimensional video display device that displays the generated multi-view video.

[0044] In step S1, the virtual camera setting means 11 sets an initial position of a virtual camera that virtually shoots a 3D model. In step S2, the 3D model deformation means 12 deforms the 3D model in the camera coordinate system corresponding to the set position of the virtual camera. Here, the 3D model deformation means 12 deforms the 3D model according to the formulas (1) and (2) so that the 3D model is enlarged and projected onto the projection plane of the virtual camera from an enlargement start point separated from the virtual camera by the recommended viewing distance.

[0045] In step S3, the virtual camera photographing means 13 virtually photographs the 3D model deformed in step S2 from the position of the virtual camera set in step S1 or step S5 described later. As a result, one viewpoint video is generated for one position of the virtual camera. Note that since the 3D model is deformed in step S2, the viewpoint video can be generated by a single photographing using orthographic projection for the photographing by this virtual camera.

[0046] In step S4, the virtual camera setting means 11 determines whether photographing has been completed at the positions of all the virtual cameras for the number of viewpoint videos in the display device configuration data. Here, if photographing has not yet been completed at the positions of all the virtual cameras (No in step S4), in step S5, the virtual camera setting means 11 sets the position of the virtual camera to the position of the virtual camera corresponding to another viewpoint video.

[0047] On the other hand, if photographing has been completed at the positions of all the virtual cameras (Yes in step S4), in step S6, the multi-viewpoint video output means 14 outputs the plurality of viewpoint videos sequentially generated in step S3 as a multi-viewpoint video. By the above operations, the multi-viewpoint video generation device 1 can generate a viewpoint video by a single photographing according to the position of the virtual camera from the CG 3D model and generate a multi-viewpoint video.

[0048] As described above, the configuration and operation of the multi-viewpoint video generation device 1 according to the embodiment of the present invention have been described, but the present invention is not limited to this embodiment. Here, as shown in FIG. 2, the device for displaying a three-dimensional video using a multi-viewpoint video is a device (three-dimensional video display device 2) in which the principal ray of the video light of each individual viewpoint video i in the multi-viewpoint video I is perpendicular to the diffusing screen 23.

[0049] However, as shown in FIG. 6, an apparatus for displaying a three-dimensional image using a multi-viewpoint image may be an apparatus (three-dimensional image display apparatus 2B) in which the principal ray of the image light of each viewpoint image i in the multi-viewpoint image I is inclined with respect to the diffusion screen 23, and the image light of the viewpoint image i is irradiated onto the same region of the diffusion screen 23. Note that the three-dimensional image display apparatus 2B is different from the three-dimensional image display apparatus 2 in that a viewing field control lens 24 is provided in front of the aperture array 22.

[0050] The viewing field control lens 24 is composed of an optical lens such as a convex lens, and irradiates the image light of each viewpoint image i onto the same region of the diffusion screen 23 spaced apart by a focal length f2. The viewing field control lens 24 may be disposed in close contact with the aperture array 22. In that case, the 3D model deformation means 12 and the virtual camera photographing means 13 may perform the following processing. That is, the 3D model deformation means 12 makes the principal ray having the inclination of the image light of the viewpoint image coincide with the z-axis of the virtual camera, and deforms the 3D model along the direction of the principal ray (z-axis). Then, the virtual camera photographing means 13 may photograph the 3D model by performing an oblique projection transformation with the virtual camera along the direction of the principal ray (z-axis).

[0051] FIG. 7 is a diagram modeling the positional relationship among the image light of the viewpoint image of the three-dimensional image display apparatus 2B in FIG. 6, the 3D model, and the virtual camera. The magnification start point Pi in FIG. 7 corresponds to the center of the aperture Ai in FIG. 6. Also, the magnification start plane P in FIG. 7 corresponds to the aperture array 22 in FIG. 6. In FIG. 7, illustration of the light ray from the viewpoint image i to the magnification start point Pi is omitted.

[0052] In the positional relationship shown in FIG. 7, the virtual camera C is disposed at a position on the extension line passing through the center Pc of the overlapping irradiation region of the viewpoint image light Lti on the diffusion screen 23 from the magnification start point Pi and spaced apart from the magnification start plane P by a viewing distance (recommended viewing distance) L.

[0053] At this time, the chief ray of the image light of each perspective image i is inclined with respect to the diffusion screen 23. The x-axis, y-axis, and z-axis are the camera coordinate system of the virtual camera C, and the x-axis and z-axis are an oblique coordinate system. That is, the xy plane is parallel to the start plane of magnification P, but the z-axis is a straight line inclined with respect to the diffusion screen 23 that connects the imaging center Ci of the virtual camera C and the start point Pi of magnification of the corresponding perspective image, and the positive direction of the z-axis is the front direction of the virtual camera C. The imaging center Ci of the virtual camera C is at the origin coordinates (0, 0, 0). In the positional relationship shown in FIG. 7, the perspective image light Lti expands from the start point of magnification Pi on the z-axis at a preset angle and is projected onto the virtual camera C's imaging center Ci on the xy plane with a width dx in the x direction and a width dy (not shown) in the y direction.

[0054] Also, the 3D model deformation means 12 deforms the 3D model M according to the positional relationship shown in FIG. 7 by the above formulas (1) and (2). Also, the virtual camera imaging means 13 generates a perspective image by obliquely imaging the deformed 3D model M along the z-axis from the position of the virtual camera C. That is, the virtual camera imaging means 13 may project the 3D model onto the imaging plane of the virtual camera C by oblique projection transformation along the z-axis for the 3D CG scene seen from the virtual camera C.

[0055] Accordingly, even when the chief ray of the light ray of the perspective image is inclined in an oblique direction, the multi-perspective image generation device 1 can generate a perspective image by a single imaging according to the position of the virtual camera, and can generate multi-perspective images more simply and at a higher speed than before.

Explanation of Reference Numerals

[0056] 1 Multi-perspective image generation device 10 Storage means 11 Virtual camera setting means 12 3D model deformation means 13 Virtual camera imaging means 14 Multi-perspective image output means C Virtual camera I Multi-view video i Viewpoint video T 3D video P Enlargement start surface Pi Enlargement start point

Claims

1. A multi-viewpoint video generating apparatus that generates the multi-viewpoint video used in a three-dimensional video display apparatus that displays a three-dimensional video by condensing the video light of a plurality of viewpoint videos of a multi-viewpoint video, expanding the video light starting from the condensation point as an expansion starting point, and superimposing the video light on a diffusion screen, comprising: 3D model deformation means for deforming the 3D model so that it is photographed by the corresponding virtual camera by expanding the 3D model in a direction parallel to the imaging plane of the virtual camera in the camera coordinate system of each of a plurality of virtual cameras virtually arranged at a predetermined viewing distance from the expansion starting point corresponding to the plurality of viewpoint videos; Virtual camera photographing means for virtually photographing the 3D model deformed by the 3D model deformation means for each of the camera coordinate systems of the plurality of virtual cameras as a viewpoint video with each virtual camera; Multi-viewpoint video output means for outputting the plurality of viewpoint videos photographed by the virtual camera photographing means as the multi-viewpoint video; A multi-viewpoint video generating apparatus, characterized by comprising the above.

2. Taking the shooting center of the virtual camera as the origin of the camera coordinate system, the shooting plane as the xy coordinate, the coordinate in the depth direction with the direction from the origin of the camera coordinate system to the start point of enlargement as positive as the z coordinate, the viewing distance as L, and the coordinates of the constituent points constituting the 3D model as (x ｉ , y ｉ , z ｉ ), when the 3D model deformation means, the xy coordinates of the constituent points constituting the 3D model are 【Number 1】 The multi-viewpoint video generating apparatus according to claim 1, characterized in that conversion is performed as described above.

3. The three-dimensional video display apparatus is configured such that the principal rays of the video light of individual viewpoint videos are perpendicular to the diffusion screen, and each viewpoint video is irradiated so as to partially overlap with each other. The virtual camera photographing means is configured to photograph the 3D model deformed by the 3D model deformation means by orthographic projection conversion. The multi-viewpoint video generating apparatus according to claim 2.

4. The three-dimensional video display apparatus is configured such that the principal rays of the video light of individual viewpoint videos are inclined with respect to the diffusion screen, and each viewpoint video irradiates the same area in an overlapping manner. The z-axis of the coordinate system of the virtual camera passes through the expansion starting point and the center of the same area. The virtual camera photographing means is configured to photograph the 3D model deformed by the 3D model deformation means by oblique projection conversion. The multi-viewpoint video generating apparatus according to claim 2.

5. A multi-viewpoint video generating program for causing a computer to function as the multi-viewpoint video generating apparatus according to any one of claims 1 to 4.

Images