Sony Patent | Image Processing Apparatus And Method, File Generation Apparatus And Method, And Program

Patent: Image Processing Apparatus And Method, File Generation Apparatus And Method, And Program

Publication Number: 20200059635

Publication Date: 20200220

Applicants: Sony

Abstract

There is provided an image processing apparatus and method, a file generation apparatus and method, and a program that enable a suitable occlusion image to be obtained. The image processing apparatus includes an MPD file processing unit configured to select an occlusion image to be acquired, on the basis of information regarding a viewpoint position of the occlusion image included in an MPD file, from among a plurality of the occlusion images indicated by the MPD file. The present technology can be applied to a client device.

TECHNICAL FIELD

[0001] The present technology relates to an image processing apparatus and method, a file generation apparatus and method, and a program, and particularly relates to an image processing apparatus and method, a file generation apparatus and method, and a program that are made capable of obtaining a suitable occlusion image.

BACKGROUND ART

[0002] Conventionally, a full spherical image is known that is an omnidirectional image (video) of 360 degrees in each direction, that is, a horizontal direction and a vertical direction.

[0003] In reproduction of the full spherical image, it is possible to experience viewing of an image (video) viewed from the center of the full sphere, by using only texture information. Furthermore, using depth information in addition to the texture information to reproduce the full spherical image can realize stereoscopic vision and motion parallax of a video. This operation allows movement of a subject to become closer to a real movement, and can improve realistic feeling to prevent motion sickness. Hereinafter, in particular, image information including the texture information and the depth information of the full spherical image is also referred to as a main image.

[0004] Furthermore, in reproduction of the full spherical image, in addition to the experience of looking around from the center (hereinafter also referred to as an origin O) of the full sphere, it is possible to perform display of looking-in in which a viewpoint position of a user as a viewer/listener is moved from the origin O, by using an occlusion image in addition to the main image.

[0005] Here, the occlusion image is image information including: texture information of an occlusion region, which is a region of the subject that does not exist in the texture information of the main image with viewpoint position being the origin O, that is, a subject not visible from the origin O; and depth information indicating a position (distance) in a depth direction regarding the occlusion region.

[0006] For example, in moving picture experts group phase-dynamic adaptive streaming over HTTP (MPEG-DASH), it is possible to transmit the main image and the occlusion image as described above to a client device (see, for example, Non-Patent Document 1).

[0007] In particular, in the MPEG-DASH, a plurality of occlusion images can be transmitted, and the client device can realize display of looking-in by using a suitable occlusion image according to a viewpoint position of the user among the plurality of the acquired occlusion images.

CITATION LIST

Non-Patent Document

[0008] Non-Patent Document 1: ISO/IEC 23009-1 Information technology-Dynamic adaptive streaming over HTTP (DASH)-Part 1: Media presentation description and segment formats, April 2012

SUMMARY OF THE INVENTION

Problems to be Solved by the Invention

[0009] However, with the above-described technology, there has been a case where it is not possible to obtain a suitable occlusion image on the client device.

[0010] For example, in the MPEG-DASH, when a transmission band is not sufficient, the client device acquires either the main image or only the texture information of the main image, without acquiring the occlusion image.

[0011] In this case, even when the transmission band has an allowance for acquiring only some occlusion image to be used for displaying the full spherical image among the plurality of occlusion images, the client device cannot acquire the occlusion image. Therefore, there has been a case where the client device cannot acquire a necessary occlusion image depending on a state of the transmission band, and cannot perform display of looking-in.

[0012] The present technology has been made in view of such a situation, and makes it possible to obtain a suitable occlusion image.

Solutions to Problems

[0013] The image processing apparatus according to a first aspect of the present technology includes an MPD file processing unit configured to select an occlusion image to be acquired, on the basis of information regarding a viewpoint position of an occlusion image included in an MPD file, from among a plurality of the occlusion images indicated by the MPD file.

[0014] An image processing method or program according to the first aspect of the present technology includes a step of selecting an occlusion image to be acquired, on the basis of information regarding a viewpoint position of the occlusion image included in an MPD file, from among a plurality of the occlusion images indicated by the MPD file.

[0015] In the first aspect of the present technology, on the basis of information regarding a viewpoint position of the occlusion image included in an MPD file, the occlusion image to be acquired is selected from among a plurality of the occlusion images indicated by the MPD file.

[0016] A file generation apparatus according to a second aspect of the present technology includes: an MPD file generation unit configured to generate an MPD file including information regarding a viewpoint position of an occlusion image for a plurality of the occlusion images; and a communication unit to send the MPD file.

[0017] A file generation method or program according to the second aspect of the present technology includes the steps of: generating an MPD file including information regarding a viewpoint position of an occlusion image for a plurality of the occlusion images; and sending the MPD file.

[0018] In the second aspect of the present technology, an MPD file including information regarding a viewpoint position of an occlusion image is generated for a plurality of the occlusion images, and the MPD file is sent.

Effects of the Invention

[0019] According to the first and second aspects of the present technology, a suitable occlusion image can be obtained.

[0020] It is to be noted that the effects described herein are not necessarily limited, and any of the effects described in the present disclosure is possible.

BRIEF DESCRIPTION OF DRAWINGS

[0021] FIG. 1 is a view for explaining occlusion.

[0022] FIG. 2 is a view for explaining an occlusion image.

[0023] FIG. 3 is a view for explaining a position of an occlusion image.

[0024] FIG. 4 is a view showing an example of an MPD file.

[0025] FIG. 5 is a view showing an example of the MPD file.

[0026] FIG. 6 is a diagram showing a configuration example of a file generation apparatus.

[0027] FIG. 7 is a flowchart for explaining an upload process.

[0028] FIG. 8 is a diagram showing a configuration example of a client device.

[0029] FIG. 9 is a flowchart for explaining a selection order list generation process.

[0030] FIG. 10 is a flowchart for explaining a file acquisition process.

[0031] FIG. 11 is a flowchart for explaining a reproduction process.

[0032] FIG. 12 is a view for explaining occlusion direction information.

[0033] FIG. 13 is a table showing an example of a selection information flag and a selection information value.

[0034] FIG. 14 is a view showing an example of the MPD file.

[0035] FIG. 15 is a flowchart for explaining the selection order list generation process.

[0036] FIG. 16 is a table for explaining notation of the occlusion direction information.

[0037] FIG. 17 is a table showing an example of the selection information flag and the selection information value.

[0038] FIG. 18 is a flowchart for explaining the selection order list generation process.

[0039] FIG. 19 is a table showing an example of the selection information flag and the selection information value.

[0040] FIG. 20 is a table showing an example of the selection information flag and the selection information value.

[0041] FIG. 21 is a table showing an example of the selection information flag and the selection information value.

[0042] FIG. 22 is a view for explaining signaling of Quality information.

[0043] FIG. 23 is a view for explaining signaling of the Quality information.

[0044] FIG. 24 is a view for explaining signaling of the Quality information.

[0045] FIG. 25 is a view showing an example of the MPD file.

[0046] FIG. 26 is a flowchart for explaining the selection order list generation process.

[0047] FIG. 27 is a diagram showing a configuration example of a computer.

MODE FOR CARRYING OUT THE INVENTION

[0048] Hereinafter, embodiments to which the present technology is applied will be described with reference to the drawings.

First Embodiment

[0049]

[0050] The present technology is to enable a client device to obtain some suitable occlusion image even in a case where a transmission band is not sufficient for transmitting all the occlusion images, in distributing a plurality of occlusion images in addition to a main image in distribution of a full spherical image using MPEG-DASH.

[0051] In particular, here, it becomes possible to select a suitable occlusion image on the client device by simultaneously signaling coordinate information indicating a viewpoint position of the occlusion image and information regarding a depth level of an occlusion region, to the client device.

[0052] Note that, in the following, a description is given to an example in which a media presentation description (MPD) file includes: information for selecting an occlusion image, such as coordinate information indicating a viewpoint position of the occlusion image; and information regarding a depth level of the occlusion region.

[0053] However, without limiting to a method of signaling information for selecting an occlusion image with the MPD file, information for selecting an occlusion image may be signaled to the client device in any other method. For example, the server or the like may supply only information for selecting an occlusion image to the client device separately from the MPD file.

[0054] Now, the present technology will be described below. Note that a description is given on assumption that a full spherical image is a moving image here, but the full spherical image may be a still image. Furthermore, the image to be processed in the present technology is not limited to an omnidirectional image, but may be some image of an omnidirectional image and the like, such as a hemispherical image.

[0055] First, a main image includes texture information, which is a moving image as a full spherical image, and depth information of the full spherical image.

[0056] The full spherical image reproduced by the texture information of the main image is to be an omnidirectional image and the like of, for example, 360 degrees in each direction, that is, in a horizontal direction and a vertical direction viewed from the origin O, in a predetermined coordinate system (hereinafter, also referred to as a full spherical coordinate system).

[0057] Here, at a start of reproduction of the full spherical image, a position of the origin O is a center position of a full sphere, which is a virtual screen of the full spherical image, and a viewpoint position of a user who views (watches) the full spherical image, that is, a center position of the user’s head, is to be at the origin O.

[0058] Furthermore, the depth information of the main image is information indicating a distance from the origin O to a position of a subject in each region on the full spherical image (texture information), that is, information indicating a position in a depth direction of each subject of the full spherical image.

[0059] Next, looking-in display and the occlusion image will be described.

[0060] For example, as indicated by arrow A11 in FIG. 1, there are a cylinder H11 and a cube H12 as subjects of the full spherical image in a space, and an image obtained by image-capturing the cylinder H11 and the cube H12 from a direction indicated by arrow V11 is to be the texture information of the main image. Furthermore, the texture information obtained by the image capturing and depth information to be obtained for the texture information are to be encoded as the main image.

[0061] In this case, as the texture information of the main image, a texture image PV11 shown in the lower left in the figure is obtained as the texture information of the main image.

[0062] Now, it is assumed that a user viewing the cylinder H11 and the cube H12 from the origin O, that is, in the direction of arrow V11 slightly changes the viewpoint position to a position on the left side in the figure from the origin O, and views the cylinder H11 and the cube H12 so as to look-in from a direction of arrow V12, that is, from the left side.

[0063] In this case, it is considered to generate, on the basis of the main image, a full spherical image viewed from the viewpoint after movement of the user, that is, viewed from the direction of arrow V12.

[0064] For example, an image (texture information) when the cylinder H11 and the cube H12 are viewed from the direction of arrow V12 is a texture image PV12 shown in the lower right in the figure. That is, in a case where the user looks in from the left side and views the cylinder H11 and the cube H12, those subjects should appear as shown in the texture image PV12.

[0065] In the texture image PV12, a hatched region R11 is a region including information of a subject that cannot be obtained from the texture image PV11, which is texture information of the main image, and this region is the occlusion region.

[0066] In other words, the region R11 on the texture image PV12 is a region that cannot be seen behind another subject or the like in the direction of arrow V11, and does not exist on the texture image PV11.

[0067] As described above, the information of the region R11, which is the occlusion region, is not included in the texture image PV11, which is the texture information of the main image, and in the depth information of the main image.

[0068] Therefore, when the texture image PV12 viewed from the direction indicated by arrow V12 is generated by using only the main image, the information of a portion of the region R11, which is the occlusion region, is lost. Therefore, for example, the client device can display the occlusion region in black, or predict pixel values of pixels in the occlusion region by using pixel information in the vicinity of the occlusion region, but the image obtained in such a manner is to be an image in which the actual subject does not appear, which causes a sense of discomfort.

[0069] Therefore, by generating, as an occlusion image, an image of texture information and depth information including information of at least a portion of the occlusion region, it becomes possible to correctly display the occlusion region that is to be lost only with the main image, by using not only the main image but also the occlusion image. With this operation, it is possible to display an image with high realistic feeling even when performing looking-in display, that is, when displaying a full spherical image with a position different from the origin O as the viewpoint position of the user.

[0070] For example, in the example shown in FIG. 1, when a viewpoint position corresponding to the direction indicated by arrow V12 is the viewpoint position of the occlusion image (hereinafter also referred to as an origin O’), the texture image PV12 itself can be taken as the texture information of the occlusion image.

[0071] In such a case, the occlusion image is to be image information including: the texture image PV12 as texture information; and depth information indicating a distance in a depth direction corresponding to the subject at each position on the texture image PV12.

[0072] Note that the texture information of the occlusion image may be an omnidirectional full spherical image when the viewpoint position is the origin O’, such as the texture image PV12, or may be an image of only the portion of the region R11 that is an occlusion region. That is, the texture information of the occlusion image is only required to include at least image information (texture information) of the occlusion region.

[0073] Furthermore, the origin O’, which is the viewpoint position of the occlusion image, may be the same as the origin O of the full spherical coordinate system in some cases.

[0074] For example, as shown in FIG. 2, it is assumed that there are an object OB1, an object OB2, and an object OB3 in a space.

[0075] In this example, when viewed from the origin O of the full spherical coordinate system, a partial region of the object OB2 is hidden by the object OB1 and invisible, and a partial region of the object OB3 is hidden by the object OB1 and the object OB2 and invisible.

[0076] In such a case, for example, the texture information of the main image is an image captured with the origin O as the viewpoint position in a state where the objects OB1 to OB3 are present.

[0077] Therefore, the texture information of the obtained main image is in a state where, although the objects OB1 to OB3 appear as subjects, a partial region of the object OB2 and a partial region of the object OB3 are hidden and invisible.

[0078] Furthermore, an image obtained by image capturing with the origin O as the origin O’ in a state where the object OB1 is absent is to be texture information of an occlusion image 1.

[0079] In this case, the texture information of the occlusion image 1 is in a state where the object OB1 does not appear as a subject, and the entire region of the object OB2 and a partial region of the object OB3 are visible. In other words, a partial region of the object OB3 is in a state of being hidden by the object OB2 and invisible.

[0080] The texture information of such an occlusion image 1 includes, as occlusion region information, information of a partial region of the object OB2 that is not included in the texture information of the main image.

[0081] Moreover, an image obtained by image capturing with the origin O as the origin O’ in a state where the object OB1 and the object OB2 are absent is to be texture information of an occlusion image 2.

[0082] In this case, the texture information of the occlusion image 2 is in a state where the object OB1 and the object OB2 do not appear as subjects, and the entire region of the object OB3 is visible.

[0083] The texture information of such an occlusion image 2 includes, as occlusion region information, information of a partial region of the object OB3 that is not included in the texture information of the main image and the occlusion image 1.

[0084] Note that, in more detail, the texture information of these occlusion image 1 and occlusion image 2 is generated on the basis of an image obtained by image-capturing the objects OB1 to OB3 with a plurality of mutually different positions as viewpoints, such as the origin O and another position different from the origin O. That is, the texture information of the occlusion image is generated on the basis of a plurality of images having mutually different viewpoint positions.

[0085] Meanwhile, there may be a case where there is a plurality of occlusion images for one main image.

[0086] As such an example, it may be considered a case where, for example, there is a plurality of occlusion images whose origin O’ is each of a plurality of positions different from the origin O of the main image in the full spherical coordinate system. In this case, the plurality of these occlusion images includes, for example, information of a region that is invisible as a subject in the main image.

[0087] Furthermore, as another example in which there is a plurality of occlusion images for one main image, it may be considered a case where there is a plurality of occlusion images having the origin O’ at a same position in the full spherical coordinate system.

[0088] In other words, for example, it may be considered a case where there is a plurality of occlusion images whose origin O’ is a position of the origin O of the main image, and these occlusion images include information of a region that is an occlusion region in the main image. At this time, for example, as described with reference to FIG. 2, some occlusion images may include information of a region that is an occlusion region in another occlusion image.

[0089] In a case where there is a plurality of occlusion images for the main image, for example, operation in the following use cases are assumed.

[0090] Use Case U1

[0091] A case where there is only an occlusion image whose origin O’* is a position of the origin O of the main image*

[0092] Use Case U2

[0093] A case where there is a plurality of occlusion images whose origin O’ is a position other than the origin O of the main image, and the origins O’* of individual occlusion images are at mutually different positions*

[0094] Use Case U3

[0095] A case where there are an occlusion image whose origin O’ is a position of the origin O of the main image and an occlusion image whose origin O is a position other than the origin O’,* and there is a plurality of occlusion images only at the position of the origin O*

[0096] Use Case U4

[0097] A case where there are an occlusion image whose origin O’ is a position of the origin O of the main image and an occlusion image whose origin O is a position other than the origin O’,* and there is a plurality of occlusion images at each of the position of the origin O and the position different from the origin O*

[0098] For example, in the use case U1, an occlusion image may not exist at a position other than the origin O, and a plurality of occlusion images may exist at the position of the origin O. Furthermore, in the use case U2, an occlusion image does not exist at the position of the origin O.

[0099] Moreover, for example, the use case U3 is a use case where the use case U1 and the use case U2 are combined.

[0100] As an example of these use cases U1 to U4, for example, an example shown in FIG. 3 can be considered. Note that, in FIG. 3, an x-axis, a y-axis, and a z-axis indicate axes of the full spherical coordinate system. Furthermore, although not shown here, the main image exists at the origin O of the full spherical coordinate system.

[0101] Moreover, a circle “o” in FIG. 3 indicates that there is one occlusion image whose viewpoint position (origin O’) is a position of the circle. A double circle “(” indicates that there are two occlusion images whose origin O’ is a position of the double circle.

[0102] For example, in the example indicated by arrow A21, two occlusion images exist at the position of the origin O at which the main image exists, and this example is the use case U1 described above.

[0103] Furthermore, in the example indicated by arrow A22, one occlusion image exists at each of a plurality of positions different from the origin O, and this example is the use case U2. In particular, here, the occlusion images exist at a plurality of positions on each axis of the x-axis, y-axis, and z-axis.

[0104] In the example indicated by arrow A23, two occlusion images exist at the position of the origin O, and one occlusion image exists at each of a plurality of positions different from the origin O on each axis of the x-axis, y-axis, and z-axis. That is, in the example indicated by arrow A23, there are both the occlusion image of the example indicated by arrow A21 and the occlusion image of the example indicated by arrow A22, and this example is the use case U3.

[0105] In the example indicated by arrow A24, two occlusion images exist at the position of the origin O, and two occlusion images also exist at each of a plurality of positions different from the origin O on each axis of the x-axis, y-axis, and z-axis. The example indicated by arrow A24 is the use case U4.

[0106] Now, consider a case of distributing the main image and the occlusion image to the client device by using MPEG-DASH.

[0107] As a specific example, for example, in the use case U1 described above, consider a case where there are two occlusion images at the origin O, and assume that there are the occlusion image 1 and the occlusion image 2 whose origin O’ is the origin O, which is a viewpoint position.

[0108] Here, the occlusion image 1 is image information including information of a region that is an occlusion region in the main image. Furthermore, the occlusion image 2 is image information including information of a region that is an occlusion region in the main image and is also an occlusion region in the occlusion image 1.

[0109] As described above, the main image includes the texture information and the depth information of the main image, and similarly, the occlusion image 1 and the occlusion image 2 also include texture information and depth information of these occlusion images.

[0110] A file for obtaining the main image having such an occlusion image 1 and an occlusion image 2, that is, an MPD file including metadata of the main image, is to be, for example, as shown in FIG. 4.

[0111] FIG. 4 shows an example in the use case U1. In FIG. 4, information regarding the texture information (texture information) of the main image is described in a portion indicated by arrow Q11, while information regarding the depth information (depth information) of the main image is described in a portion indicated by arrow Q12.

[0112] Furthermore, information regarding the texture information of the occlusion image 1 is described in a portion indicated by arrow Q13, while information regarding the depth information of the occlusion image 1 is described in a portion indicated by arrow Q14.

[0113] Moreover, information regarding the texture information of the occlusion image 2 is described in a portion indicated by arrow Q15, while information regarding the depth information of the occlusion image 2 is described in a portion indicated by arrow Q16.

[0114] The portion indicated by each of arrows Q1 to 016 is one piece of AdaptationSet.

[0115] For example, in the portion indicated by arrow Q11, “vt1” is described as id of Representation of the texture information of the main image.

[0116] Similarly, in the portion indicated by arrow Q13, “vot1” is described as id of Representation of the texture information of the occlusion image 1. Further, in the portion indicated by arrow Q15, “vot2” is described as id of Representation of the texture information of the occlusion image 2.

[0117] Furthermore, in the portion indicated by arrow Q12, a portion indicated by arrow Q14, and the portion indicated by arrow Q16, which are portions related to the depth information, a portion of schemeIdUri in EssentialProperty of the MPEG-DASH is to be schemeIdUri=”urn:mpeg:dash:depth:2015”.

[0118] That is, in AdaptationSet of the depth information of each of the main image, the occlusion image 1, and the occlusion image 2, schemeIdUri is set to “urn:mpeg:dash:depth:2015”. These descriptions of schemeIdUri allow the portion indicated by arrow Q12, the portion indicated by arrow Q14, and the portion indicated by arrow Q16 to be specified as being a description related to the depth information.

[0119] Furthermore, in Representation of the depth information, associationId is used to describe id of the texture information corresponding to the depth information, which makes it possible to specify which texture information the depth information corresponds to.

[0120] For example, the portion indicated by arrow Q12 describes associationId=”vt1”, and indicates a relation (association) between the texture information of the main image in the portion indicated by arrow Q11 and the depth information of the main image in the portion indicated by arrow Q12.

[0121] Similarly, for example, in the portion indicated by arrow Q14, id “vot1” described in the portion indicated by arrow Q13 is used to describe associationId=”vot1”, while in the portion indicated by arrow Q16, id “vot2” described in the portion indicated by arrow Q15 is used to describe associationId=”vot2”.

[0122] With these descriptions, the texture information and the depth information can be correctly specified for the main image, the occlusion image 1, and the occlusion image 2.

[0123] Moreover, whether the texture information is associated with the occlusion image is made distinguishable by setting the portion of schemeIdUri as “urn:mpeg:dash:occlusion:2015” in EssentialProperty, for example, as described in the portion indicated by arrow Q13 and the portion indicated by arrow Q15.

[0124] That is, schemeIdUri is set to “urn:mpeg:dash:occlusion:2015” in AdaptationSet of the texture information of each of the occlusion image 1 and the occlusion image 2. These descriptions of schemeIdUri allow the portion indicated by arrow Q13 and the portion indicated by arrow Q15 to be specified as being a description related to the texture information of the occlusion image.

[0125] Using the MPD file including the information regarding the occlusion image as described above enables the following implementation in client device in the MPEG-DASH. In other words, the following processes TR1 to TR3 can be executed.

[0126] Process TR1

[0127] All of the main image, the occlusion image 1, and the occlusion image 2** are acquired to reproduce a full spherical image**

[0128] Process TR2

[0129]* Only the texture information and the depth information of the main image are acquired to reproduce a full spherical image*

[0130] Process TR3

[0131]* Only the texture information of the main image is acquired to reproduce a full spherical image*

[0132] In this case, for example, in the process TR2, stereoscopic vision and motion parallax can be realized since the depth information of the main image is acquired, but an occlusion region cannot be correctly displayed when looking-in display is performed, since the occlusion image is not acquired.

[0133] Furthermore, in the process TR3, stereoscopic vision and motion parallax cannot be realized since the depth information of the main image is not acquired.

[0134] In a case where a transmission bandwidth between the self and a server that distributes the main image and the like is not sufficient, and all of the main image, the occlusion image 1, and the occlusion image 2 cannot be acquired from the server, the client device is to select either the process TR2 or the process TR3.

[0135] Specifically, the client device first measures the transmission band, and selects the texture information of the main image as the information to be acquired from the server.

[0136] Next, when there is a transmission band for acquiring the depth information of the main image on the basis of a measurement result of the transmission band, the client device also selects the depth information of the main image as the information to be acquired.

[0137] On the other hand, when the transmission band is not sufficient, the client device is to acquire only the texture information of the main image.

[0138] Furthermore, in a case where the client device selects the texture information and the depth information of the main image as the information to be acquired, it is determined whether or not there is a sufficient transmission band for acquiring the occlusion image 1 and the occlusion image 2 on the basis of the measurement result of the transmission band.

[0139] Then, in a case where it is determined that there is a sufficient transmission band, the client device selects, as the information to be acquired, the texture information and the depth information of each image of the main image, the occlusion image 1, and the occlusion image 2 as the information to be acquired from the server.

[0140] Furthermore, in a case where it is determined that there is not a sufficient transmission band, the client device selects, as the information to be acquired, the texture information and the depth information of the main image as the information to be acquired from the server.

[0141] When the information to be acquired from the server is this selected, the client device acquires the selected information from the server, and reproduces a full spherical image. With this operation, any one of the processes TR1 to TR3 described above is to be performed.

[0142] In particular, in a case where there is not a sufficient transmission band for acquiring the occlusion image 1 and the occlusion image 2, either the process TR2 or the process TR3 is performed.

[0143] However, depending on the transmission band, it may be possible to transmit either one of the occlusion image 1 or the occlusion image 2 in addition to the main image.

[0144] Furthermore, at a time of reproduction of the full spherical image, it is possible to use only some occlusion image among a plurality of occlusion images.

[0145] For example, if some occlusion image of the plurality of occlusion images includes occlusion region information necessary for displaying the full spherical image, it is only required to use only the occlusion image including the necessary information regarding the occlusion region.

[0146] Therefore, it is useful to selectively acquire and use only some occlusion image among a plurality of occlusion images.

[0147] In this example, since the occlusion image 2 includes occlusion region information that is not included in the occlusion image 1 as well, the occlusion image 1 is more useful in the occlusion image 1 and the occlusion image 2.

[0148] In other words, for example, it is assumed that the occlusion image 1 mentioned here is the occlusion image 1 described with reference to FIG. 2, and the occlusion image 2 is the occlusion image 2 described with reference to FIG. 2.

[0149] In this example, the occlusion image 1 includes occlusion region information of the subject (object OB2) further on the front side (the origin O side) than a case of the occlusion image 2. Therefore, it can be said that it is effective to use the occlusion image 1 before the occlusion image 2 at a time of full spherical image reproduction. In other words, it can be said that the occlusion image 1 is a more suitable occlusion image to be acquired (used) than the occlusion image 2 out of the two occlusion images.

[0150] However, it is difficult for the client device to select (specify) which occlusion image of the occlusion image 1 and the occlusion image 2 is more suitable, that is, which has a higher priority. This is because the client device cannot obtain information for selecting a suitable occlusion image, in a current state.

[0151] Not only in the use case U1 described above, but also in the use case U2, the use case U3, and the use case U4, in a case where there are a large number of occlusion images, it is similarly difficult to select suitable one or more occlusion images from among these occlusion images.

[0152] Therefore, the client device has been unable to obtain a suitable occlusion image when the transmission band is not sufficient.

[0153] Therefore, in the present technology, it is made possible to select and acquire some suitable occlusion image from among a plurality of occlusion images, and consequently made possible to use a transmission band more efficiently.

[0154] Specifically, in the present technology, as information for selecting an occlusion image, occlusion coordinate information indicating a position of a viewpoint position of the occlusion image, that is, a position of the origin O’, in the full spherical coordinate system and depth level information indicating a depth level of the occlusion image are to be used.

[0155] Here, the depth level information (hereinafter also described as depth level) is information based on a distance (depth) from the origin O’ indicated by the depth information of the occlusion image, to the subject in the occlusion image. In other words, the depth level information is information regarding a distance (depth) in a depth direction of the subject in the occlusion image.

[0156] By using the occlusion coordinate information and the depth level information in this manner, a suitable occlusion image to be acquired from the server can be selected in accordance with a viewpoint position of a user, even in a case where there are occlusion images at a plurality of positions in the full spherical coordinate system, for the main image.

[0157] Specifically, for example, it is assumed that, in the use case U2 described above, that is, for example, in a case where there is an occlusion image at the position indicated by arrow A22 in FIG. 3, a viewpoint position of the user is at a position in a positive direction on the x-axis of the full spherical coordinate system, and a full spherical image viewed from the viewpoint position of the user is displayed.

[0158] At this time, the occlusion image at the position in the positive direction on the x-axis of the full spherical coordinate system includes occlusion region information necessary for generating the full spherical image for display.

[0159] Whereas, an occlusion image at a position in a negative direction on the x-axis of the full spherical coordinate system does not include occlusion region information necessary for displaying the full spherical image according to the viewpoint position of the user.

[0160] Furthermore, the occlusion image at a position on the y-axis or the z-axis of the full spherical coordinate system may include occlusion region information necessary for displaying the full spherical image, but an amount of the information should be small as compared to the occlusion image in the positive direction on the x-axis.

[0161] This fact shows that the occlusion image to be selected from among the plurality of occlusion images differs depending on the viewpoint position of the user.

[0162] In order to select a suitable occlusion image necessary for generating the full spherical image, it is necessary to know which position the occlusion image is at, as viewed from the position of the main image, that is, the position of the origin O.

[0163] As information that can specify which position the position (origin O’) of each occlusion image is at as viewed from the origin O, which is the position of the main image, it is only required to use the occlusion coordinate information described above.

[0164] The occlusion coordinate information is information regarding a viewpoint position of the occlusion image. For example, the occlusion coordinate information is to be coordinates of a position of the origin O’ of each occlusion image when coordinates of a position of the origin O of the main image in the full spherical coordinate system are (0, 0, 0), and a unit of coordinates in each axial direction is meter or the like.

[0165] Thus, for example, coordinates of a position at a distance of X meters in the positive direction of the x-axis from the origin O in the full spherical coordinate system, at a distance of Y meters in the positive direction of the y-axis from the origin O, and at a distance of Z meters in the positive direction of the z-axis from the origin O are (X, Y, Z).

[0166] With such occlusion coordinate information, it is possible to select a suitable occlusion image from a viewpoint position of the user and the occlusion coordinate information.

[0167] In other words, on the basis of the occlusion coordinate information, from among a plurality of occlusion images, it is only required to select, as an occlusion image to be acquired, sequentially from one in which viewpoint position of the occlusion image indicated by the occlusion coordinate information is closer to the viewpoint position of the user.

[0168] In other words, it is only required to select, as the occlusion image to be acquired, in ascending order of a distance between the viewpoint position of the occlusion image and the viewpoint position of the user.

[0169] Note that, in a case where there is a plurality of occlusion images at a same position, it is only required to select (utilize) sequentially from one in which a position of the occlusion region included in the occlusion image in the full spherical coordinate system is closer to a position of the subject in the main image. This is because, as described above, as the position of the occlusion region is closer to the position of the subject of the main image, a utility value is higher.

[0170] In order to select the occlusion image whose position of the occlusion region is closer to the position of the subject in the main image from among a plurality of occlusion images having the origin O’ at a same position, it is only required to use depth level information.

[0171] The depth level information (depth level) is information that is generated on the basis of the depth information regarding the occlusion image and is associated with a depth (depth amount) of the occlusion region in the occlusion image.

[0172] Here, it is assumed that the depth amount (depth) of the occlusion region is deeper as the occlusion region in the occlusion image is at a position farther from the origin O’.

[0173] In this case, for example, among a plurality of occlusion images at a same position and the like, the depth level information of the occlusion image is to be information indicating an order of a depth amount such as an order of being shallower of the depth amount of the occlusion region of the occlusion image.

[0174] Specifically, for example, among a plurality of occlusion images at a same position, depth level of the occlusion image whose occlusion region has a shallowest depth amount (not deep) is set to 1, and a depth level is set to 2, 3, … in an order from one having a smaller depth amount of the occlusion region.

[0175] In determining the order of the depth amount indicated by the depth level information (depth level), a sum of depth values at individual positions in a predetermined frame of the texture information of the occlusion image, that is, a sum of distances from the origin O’ in the full spherical coordinate system to the subject at that position is determined.

[0176] Then, among the plurality of occlusion images at a same position, depth level is made smaller as the occlusion image has a smaller total value of the depth values. That is, for example, depth level information (depth level) of an occlusion image with a smallest total value of depth values is set to 1.

[0177] Note that, here, an example has been described in which depth level is determined on the basis of a total value of depth values at individual positions. However, without limiting to this, the determination may be made on the basis of information indicating a depth amount (depth) of the occlusion region of each occlusion image, such as an average value of depth values at individual positions.

[0178] For example, in a case where a depth level is determined on the basis of an average value of the depth values at individual positions of the occlusion image, it is only required to set a value of depth level to be smaller, that is, set the order of the depth amount to be smaller, for an occlusion image with a smaller average value. Note that, hereinafter, information indicating the average value of the depth values at individual positions of the occlusion image is also referred to as depth average value information.

[0179] If such depth level information (depth level) is used, a suitable occlusion image can be selected on the client device side even in a case where there is a plurality of occlusion images at a same position. In other words, in a case where there is a plurality of occlusion images at a same position, it is only required to select the occlusion images sequentially from one with a smaller value of the depth level information.

[0180] If there are occlusion coordinate information and depth level information as described above, a suitable occlusion image can be selected in each of the use cases U1 to U4.

[0181] Here, a description is given to an example of a method for signaling occlusion coordinate information and depth level information to the client device in the MPEG-DASH.

[0182] Specifically, for example, in the MPD file, new EssentialProperty is defined that enables setting of occlusion coordinate information and depth level information for an occlusion image.

[0183] In other words, in the example shown in FIG. 4, in EssentialProperty for the texture information of the occlusion image, the portion of schemeIdUri has been set as “urn:mpeg:dash:occlusion:2015”.

[0184] Instead of this, in the present technology, a portion of schemeIdUri is set as “urn:mpeg:dash:occlusion:2017”, and value of schemeIdUri describes a value (x, y, z) of the occlusion coordinate information and a value of the depth level information depth level. Specifically, the values of value are separated by commas to be “x, y, z, depth level”.

[0185] Therefore, for example, when the occlusion coordinate information is (0, 0, 0) and the value of the depth level information depth level is 1, a portion of EssentialProperty is set as .

[0186] As described above, in a case where the MPD file is made to include the occlusion coordinate information and the depth level information of each occlusion image, for example, the MPD file that is applied with the present technology and corresponds to the MPD file shown in FIG. 4 is to be as shown in FIG. 5. Note that, in FIG. 5, a description of a portion corresponding to a case in FIG. 4 is to be omitted suitably since it becomes repetition.

[0187] The MPD file shown in FIG. 5 differs from the MPD file shown in FIG. 4 only in portions indicated by arrows Q21 and Q22.

[0188] In other words, as indicated by arrow Q21, in AdaptationSet of texture information of the occlusion image 1 in FIG. 5, schemeIdUri is set to “urn:mpeg:dash:occlusion:2017”, and a value of value thereof is set to “0, 0, 0, 1”.

[0189] This fact shows that occlusion coordinate information of the occlusion image 1 is (0, 0, 0), and a value of the depth level information depth level is 1.

[0190] Similarly, as indicated by arrow Q22, in AdaptationSet of the texture information of the occlusion image 2, schemeIdUri is set to “urn:mpeg:dash:occlusion:2017”, and a value of value thereof is set to “0, 0, 0, 2”. This description shows that occlusion coordinate information of the occlusion image 2 is (0, 0, 0), and a value of the depth level information depth level is 2.

[0191] Thus, in this example, it can be seen that there are two occlusion images 1 and 2 for the main image, the positions of the origin O’ of those occlusion images are both the origin O, which is the position of the main image.

[0192] Furthermore, in a case where there is only an allowance for acquiring one of these occlusion image 1 and occlusion image 2 as a result of the measurement of the transmission band, it is only required to select the occlusion image 1 having a smaller value of the depth level information since the positions of these occlusion images are the same.

[0193] As described above, according to the present technology, it is possible to select a suitable occlusion image by signaling occlusion coordinate information and depth level information of each occlusion image with the MPD file. As a result, the transmission band can be used more efficiently.

[0194]

[0195] Subsequently, a more specific embodiment to which the present technology is applied will be described.

[0196] FIG. 6 is a diagram showing a configuration example of an embodiment of a file generation apparatus to which the present technology is applied.

[0197] A file generation apparatus 11 shown in FIG. 6 includes a control unit 21 and a file generation unit 22. This file generation apparatus 11 generates a segment file and an MPD file of a moving image of a content to be distributed by the MPEG-DASH or the like, and uploads to a server. Note that the moving image of the content mentioned here is a full spherical image for display generated from the above-described main image and an occlusion image related to the main image.

[0198] The control unit 21 controls an overall operation of the file generation apparatus 11. For example, the control unit 21 controls the file generation unit 22 to generate a segment file storing a full spherical image or the like, which is a moving image of content, or an MPD file including metadata of the content, and to upload those segment file and MPD file.

[0199] The file generation unit 22 generates a segment file and an MPD file in accordance with the control of the control unit 21, and uploads (sends) those segment file and MPD file to the server via a network.

[0200] The file generation unit 22 includes a data input unit 31, a data encoding/generation unit 32, an MPD file generation unit 33, a recording unit 34, and an upload unit 35.

[0201] The data input unit 31 acquires a main image and image data necessary for generating an occlusion image related to the main image, and metadata necessary for generating an MPD file, such as occlusion coordinate information and depth level information, and supplies to the data encoding/generation unit 32 and the MPD file generation unit 33.

[0202] The data encoding/generation unit 32 generates a segment file storing the main image and the occlusion image on the basis of the image data supplied from the data input unit 31, and supplies to the recording unit 34.

[0203] The data encoding/generation unit 32 includes a pre-processing unit 41, an encoding unit 42, and a segment file generation unit 43.

[0204] On the basis of the image data supplied from the data input unit 31, the pre-processing unit 41 generates a main image and an occlusion image by performing stitch processing for connecting images, and supplies to the encoding unit 42. The encoding unit 42 encodes the main image and the occlusion image supplied from the pre-processing unit 41, and supplies the obtained coded data to the segment file generation unit 43.

[0205] On the basis of the metadata and the like supplied from the data input unit 31, the segment file generation unit 43 files the coded data supplied from the encoding unit 42 in units of segments, and supplies a segment file obtained as a result to the recording unit 34. With this operation, a segment file storing the coded data of the texture information of the main image, a segment file storing the coded data of the depth information of the main image, and a segment file storing the coded data of the occlusion image are obtained.

[0206] The MPD file generation unit 33 generates an MPD file including information regarding the main image and the occlusion image, on the basis of the metadata supplied from the data input unit 31, and supplies to the recording unit 34. Note that the MPD file generation unit 33 may be adapted to acquire metadata necessary for generating the MPD file from the segment file generation unit 43.

[0207] The recording unit 34 records the MPD file supplied from the MPD file generation unit 33 and the segment file supplied from the segment file generation unit 43.

[0208] The upload unit 35 reads the MPD file and segment file of the content from the recording unit 34, and uploads to the server. In other words, the upload unit 35 functions as a communication unit that sends the MPD file and the segment file to the server.

[0209] Note that, here, an example will be described in which the file generation apparatus 11 functions as an apparatus to upload the MPD file and the segment file to the server, but the file generation apparatus 11 may function as the server. In such a case, the upload unit 35 of the file generation apparatus 11 sends the MPD file and the segment file to the client device via the network.

[0210]

[0211] Next, an operation of the file generation apparatus 11 will be described. In other words, an upload process by the file generation apparatus 11 will be described below with reference to the flowchart of FIG. 7.

[0212] In step S11, the data input unit 31 acquires image data necessary for generating a main image and an occlusion image, and metadata such as occlusion coordinate information and depth level information, to supply to the pre-processing unit 41, the segment file generation unit 43, and the MPD file generation unit 33.

[0213] In step S12, the data encoding/generation unit 32 generates a segment file.

[0214] In other words, on the basis of the image data supplied from the data input unit 31, the pre-processing unit 41 generates a main image and an occlusion image by performing stitch processing, and supplies to the encoding unit 42. The encoding unit 42 encodes the main image and the occlusion image supplied from the pre-processing unit 41, and supplies the obtained coded data to the segment file generation unit 43.

[0215] On the basis of the metadata and the like supplied from the data input unit 31, the segment file generation unit 43 files the coded data supplied from the encoding unit 42, and supplies a segment file obtained as a result to the recording unit 34.

[0216] In step S13, the MPD file generation unit 33 generates an MPD file on the basis of the metadata supplied from the data input unit 31, and supplies to the recording unit 34.

[0217] Here, the MPD file includes occlusion coordinate information and depth level information for each occlusion image. Specifically, for example, the MPD file shown in FIG. 5 is generated.

[0218] In step S14, the recording unit 34 records the MPD file supplied from the MPD file generation unit 33 and the segment file supplied from the segment file generation unit 43.

发表评论

电子邮件地址不会被公开。 必填项已用*标注