Video Server System

Abstract

To automatically apply a water mark that has a large suppression effect to unauthorized use of an image.SOLUTION: A request of distribution of a specific video for distribution and log-in are performed (S51). An authentication condition is transmitted to an authentication server side, the authentication server determines whether a user is regularly registered and distribution is possible (S52). When the distribution is possible (S52: Yes), the authentication server notifies a video server of information to the effect that the distribution is possible, and the video server transmits the video for distribution to a user side reproduction device (S53). Also, the video server transmits mark application information corresponding to the video for distribution to the user side reproduction device (S54). On the user side reproduction device, a control unit decodes the received video for distribution, recognizes an image for each frame, and applies a water mark to the image according to a mark application condition (S55). The water mark used herein is character information that can identify the user.SELECTED DRAWING: Figure 7

Description

[Technical field]

[0001] The present invention relates to a video server system that adds a watermark to a distributed video. [Background technology]

[0002] It is desirable to use high-quality images captured by an imaging device without any degradation. However, for example, for parts of a video that are undesirable to distribute as is (such as a captured license plate number of a car or the face of a passerby who happens to be passing by), it is desirable to distribute these parts after rendering them so that they are not clearly identifiable. Patent Document 1 describes a video editing system that automatically performs such rendering processing.

[0003] On the other hand, various technologies are known to prevent unauthorized use of video (use by others without the permission of the copyright holder, etc.) For this purpose, protection technologies such as DRM technology that limits the software that can be used to play the video and (invisible) digital watermarking that adds information to the image data that cannot be seen on the image are used, but it is not easy to prevent unauthorized use of the image by using, for example, screenshots (hard copies of the screen) with these protection technologies.

[0004] In contrast to this type of processing, a process of adding a visible watermark to an image is also performed to prevent unauthorized use of the video (image). The task of adding a watermark is similar to the rendering process described in Patent Document 1, but unlike the rendering process, the purpose is not to eliminate the distinctiveness of a part of the image, but to prevent the image from being used without permission from the copyright holder by screenshots, etc. For this reason, the content of the added watermark is, for example, text information corresponding to the copyright holder of the image. Also, in order to maintain the distinctiveness of the original image while ensuring the distinctiveness (visibility) of the watermark, the watermark is often added by being superimposed semi-transparently on the original image.

[0005] If the use of the original image is legitimately permitted, the user will be provided with the image before the watermark was added by the copyright holder, and such an image with the watermark will not be used by the user. Therefore, a user who sees an image with a recognized watermark will recognize that the image is likely to have been stolen, and can also identify the copyright holder. Therefore, the use of such watermarks can prevent unauthorized use of images.

[0006] While the above-mentioned DRM technology and digital watermarking are not effective against screenshots, etc., watermarks are visibly added to images and are therefore effective against unauthorized use using screenshots. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] JP 2019-16381 A DISCLOSURE OF THEINVENTION [Problem to be solved by the invention]

[0008] There is a demand for a technology to automatically add watermarks, like the technology described in Patent Document 1 for rendering processing. However, the contents of watermarks and the methods of adding them that have been conventionally used have the following problems.

[0009] Generally, video data is stored and distributed as encoded data, and only users with access rights to this data can obtain and decode this data to play the video. Videos (images) played in this way are often subject to unauthorized use (piracy) by screenshots as mentioned above. For this reason, in reality, there are many routes for video piracy corresponding to the number of users.

[0010] For this reason, in order to prevent theft, it is more preferable to identify the playback route of the stolen video (image) than to identify the copyright holder with a watermark. In other words, the contents of watermarks used in the past were not sufficiently effective in preventing the unauthorized use of images.

[0011] Also, for example, the copyright holder of an image to be protected is often obvious to the person viewing the image, and in such cases, there is little point in using a watermark indicating the copyright holder as described above. On the other hand, such a playback route is generally difficult to identify from the image alone. For this reason, the watermark can be text information for identifying the playback route. However, while copyright holders are generally well-known and their display can be simplified, the text information identifying the playback route is more complicated than the text information identifying the copyright holder. For this reason, when making this into a watermark, the method of applying it on the screen becomes an issue, and it is not easy to set it up.

[0012] For this reason, there was a demand for technology that could automatically add a watermark that would have a significant deterrent effect against the unauthorized use of images.

[0013] The present invention has been made in view of the above circumstances, and has an object to solve the above problems. [Means for solving the problem]

[0014] The present invention is a video server system that distributes video for distribution stored in a recording means to a preregistered user and displays the video for distribution on the user's side with a watermark added to it, comprising: a playback means that is operated by the user, obtains the video for distribution only when the user is authenticated, and displays the video for distribution with the watermark added; and a marking information creation means that analyzes images for each frame in the video for distribution to estimate the subject in the image, recognizes the position of the subject in the image, and creates marking information that defines the position at which the watermark should be added so that the subject and the watermark overlap in the image, and the manner in which the watermark should be added, and stores the marking information in the recording means in combination with the video for distribution, the watermark including identification information that identifies the user; and the playback means that obtains the marking information when obtaining the video for distribution, and displays the video for distribution with the watermark added in accordance with the marking information. The marking information generating means may also generate a single screen The marking information may be generated so as to display the watermarks of different sizes at a plurality of different positions. The marking information creating means may create the marking information so that frames to which the watermark is not added are provided between frames to which the watermark is added in the video to be distributed. The marking information creating means may create the marking information so that the watermark is overlapped on a highly distinctive portion of the subject. The marking information creating means may create the marking information so that the watermark overlaps a predetermined portion of the subject. The marking information creating means may create the marking information so that the watermark overlaps the center of the subject. Effect of the Invention

[0015] According to the present invention, a watermark that has a significant deterrent effect against unauthorized use of an image can be automatically added. [Brief description of the drawings]

[0016] [Figure 1] 1 is a diagram showing a configuration of a video server system according to an embodiment; [Diagram 2] 1 is a diagram showing a configuration of a video server used in a video server system according to an embodiment; [Diagram 3] 5 is a flowchart showing an operation of creating marking information in the video server system according to the embodiment. [Figure 4] 10 is a flowchart showing operations for analyzing an image and determining marking conditions when creating marking information. [Diagram 5] 13 is a flowchart showing an operation for estimating a subject when creating mark-adding information. [Figure 6] 2 is a diagram showing the configuration of a user-side playback device used in the video server system according to the embodiment; FIG. [Figure 7] 5 is a flowchart showing an operation performed when distributing video to be distributed in the video server system according to the embodiment. [Figure 8] 1 shows an example (part 1) of an image after a watermark is added in the video server system according to the embodiment. [Figure 9] 13 is a second example of an image after a watermark is added in the video server system according to the embodiment. [Figure 10] 13 is a third example of an image after a watermark is added in the video server system according to the embodiment.

[0017] Next, an embodiment of the present invention will be specifically described with reference to the drawings. FIG. 1 is a diagram showing the configuration of a video server 1 according to an embodiment of the present invention. This video server system 1 distributes recorded video to a large number of users. This operation is mainly performed by a video server 10, and users who can become distribution destinations are registered in advance. The video server 10 and all the components described below are connected to a network N.

[0018] Furthermore, when this video is played back by the user at the distribution destination, a watermark is automatically added to this video. The content displayed as the watermark here is information that can identify this user (such as a user name). Therefore, the watermark added here will be different for each user to whom the video is distributed. Furthermore, the display position and display mode of this watermark within the video (image) are specified by the video server 10.

[0019] In Fig. 1, a recording device (recording means) 20 stores video to be distributed in a form encoded in, for example, MPEG-2 format, where the original material data has been cut and edited for distribution (deleting video from unnecessary time periods or inserting other video into the deleted time periods, etc.) and rendered (such as mosaic processing of parts of the video) in advance by an editing device 30. No watermark is added to the video to be distributed. A large number of videos to be distributed are actually stored, and each time one of them is selected and specified, the selected video to be distributed is distributed to the user.

[0020] The sending server 40 distributes the video to be distributed that the video server 10 has acquired from the recording device 20 to a distribution destination (user) designated by the video server 10. In reality, the distribution destination is a user-side playback device (playback means) 100 (100A-100C), which is a playback device (e.g., a PC) corresponding to each user. Here, three user-side playback devices are provided, but in reality, a larger number is set. The users U of the user-side playback devices 100A-100C are referred to as users UA-UC, respectively.

[0021] As described above, only those who have been registered in advance are authenticated as users who can operate the user-side playback device 100 in this video server system 1. For this reason, an authentication server 50 is used for authenticating users.

[0022] 2 is a diagram showing the configuration of this video server 10. Here, there is provided a control unit 11 equipped with a CPU that controls the entire video server 10. To enable the control unit 11 to perform this operation, there are provided an operation unit 12 such as a keyboard or touch panel that accepts operations from the user U, and a display unit 13 that is a display that displays various information. In addition, there is also provided a working memory 14 that is used when the control unit 11 operates.

[0023] Also provided is a storage unit 15, which is composed of a large-capacity hard disk or non-volatile memory, for storing encoded video for distribution and marking information (described later). Also provided is an interface unit 16 for exchanging various types of data with the other components in Fig. 1 via the network N.

[0024] First, we will explain the flow of video distribution in this video server system 1. There is no difference in this flow between this video server system 1 and a conventional video server system.

[0025] For example, one of the user-side playback devices 100A-100C (e.g., the user-side playback device 100A) can request the video server 10 to distribute one video for distribution stored in the recording device 20 by an operation of the user U (UA). At this time, the video server 10 can recognize that the user U (UA) who performed this operation is properly registered through various authentication processes using the authentication server 50. Therefore, the video server 10 can determine that this request is valid, read out the video for distribution in an encoded form from the recording device 20, and transmit it to the user-side playback device 100 (100A) via the transmission server 40. The user-side playback device 100 (100A) that receives this can decode it and play it as a video, and the user U (UA) can view it on a display or the like.

[0026] In this video server system 1, a watermark is added to the video played back on the user-side playback device 100, and the manner in which this watermark is added is a feature of this video server system 1. This point will be described below.

[0027] When new video for distribution is obtained to be stored in the recording device 20, the video server 10 creates marking information that specifies in advance where and in what form a watermark should be added in the video (image for each frame) of the video for distribution, and stores the marking information together with the video for distribution in the recording device 20. In other words, the video server 10 functions as a marking information creating means that creates the marking information.

[0028] Fig. 3 is a flow chart showing this operation. In Fig. 3, first, when the video server system 1 (or the recording device 20) obtains new video for distribution, the control unit 11 of the video server 10 recognizes this (S1). The above-mentioned marking information is created for this video for distribution. As mentioned above, at this stage, no watermark has been added to this video.

[0029] Next, the control unit 11 analyzes the video for distribution for each frame. For this purpose, the control unit 11 reads and recognizes the image for each frame (S2), analyzes the specific content of the image (S3), and determines the conditions for adding a watermark to this image (settings of position, size, number, color, etc.: marking conditions) according to the analysis results (S4). After this operation has been performed for all frames (S5: Yes), the control unit 11 stores marking information specifying the marking conditions as described above for all frames together with the video for distribution in the recording device 20 (S6).

[0030] A plurality of distribution videos are stored in the recording device 20, one of which is specified and distributed by the user-side playback device 100, and a watermark is added when this is played back by the user-side playback device 100. At this time, the specific content of this watermark differs depending on the user U who issued the request, so the video with the watermark added cannot be shared among all users U or user-side playback devices 100. For this reason, even when the same distribution video is selected, the task of adding a watermark to it is performed each time it is distributed. In contrast, if the marking information is stored in the recording device 20 together with the distribution video as described above, the task of adding a watermark each time it is distributed can be performed extremely easily.

[0031] Next, the image analysis (S3) and the determination of the marking conditions (S4) in Fig. 3 will be described in particular. In the image analysis (S3), the object that is the main subject in the image is estimated. The subject may be people (including celebrities), animals, various buildings, etc. that the control unit 11 can identify using image identification technology. The marking conditions (S4) are set on the premise that the subject is present in the image, so that both the subject and the watermark are easily visible.

[0032] FIG. 4 is a flow chart showing these steps in detail. This operation is performed by the control unit 11. In the analysis of the image (S3), first, an object that is estimated to be the subject in the image is determined (S31). This determination is performed using AI technology in particular. At this time, not only the image but also other information can be used at the same time.

[0033] 5 is a flowchart showing in more detail the step (S31) for estimating the subject in the image. In this operation, first, the target image is recognized (S311), and then the control unit 11 checks whether or not there is a character string in the image by a pattern recognition method (S312). If a character string is recognized (S312: Yes), the control unit 11 obtains the character string as auxiliary information for estimating the subject (S313). The number of pieces of auxiliary information is arbitrary.

[0034] Next, the control unit 11 checks whether or not there is an audio signal in the time period including this image from the data of the video to be distributed (S314). If there is an audio signal (S314: Yes), the control unit 11 obtains the audio signal as a character string converted by a voice recognition method, and obtains this as auxiliary information similar to the above (S315).

[0035] If one or more pieces of auxiliary information are obtained (S316: Yes), the control unit 11 estimates the subject based on all of the auxiliary information and the content of the image itself (S317). To this end, the control unit 11 first extracts keywords from the auxiliary information, which is text information, and on the other hand, analyzes the image using well-known face analysis technology, object analysis technology, etc. to recognize objects (including people) recognized in the image. At this time, various databases stored in the storage unit 15 or obtained via the network N can be used. If no auxiliary information is obtained (S316: No), the control unit 11 estimates the subject based only on the image itself (S318). At this time, a similar database can be used.

[0036] When the subject recognized as a person is a person, the database containing information on facial images of famous people can be used to identify the person. In this case, if the database also contains information on voices (voiceprints, etc.), the audio signal accompanying the image can be used in estimating the subject (person) (S317, S318) instead of the auxiliary information in the form of a character string (S315).

[0037] Here, the object analysis technology can utilize the characteristics unique to images captured by broadcast video cameras. Considering the purpose of filming and reporting by television broadcasting stations and the like, it is believed that video camera footage often follows and captures a specific subject in the center of the screen. It is also believed that the subject appears in many scenes in the video. For these reasons, it is possible to estimate the subject with high accuracy by taking into account the frequency and number of times that an object recognized in object analysis is located in the center of the screen.

[0038] If the above analysis method does not find any object in the image that is estimated to be the subject (for example, if there is no pattern with a contrast or size that can be recognized as a subject), for example, an area of ​​a certain size in the center of the image may be set as the subject. Since the subject here is set (estimated) only as a reference when adding a watermark as described later, if an object that can be estimated as the subject cannot be identified, a region in the image may be tentatively estimated as the subject according to a predetermined rule.

[0039] In FIG. 4, after the object to be estimated as the subject is determined as described above, control unit 11 then recognizes the color distribution of the range including this subject in the image (inside the subject and around the subject) (S32).

[0040] Next, the control unit 11 determines the position in the image to which the watermark is to be added (S41) according to the position of the subject in the image recognized as described above (S31). Here, this position is set so that the semi-transparent watermark is formed at a position in the image that overlaps with the subject or in the vicinity of the subject. Also, the size, number, angle, etc. of the watermark to be displayed are set according to the size of the subject in the image, etc. Here, the size of the watermark (length in the vertical and horizontal directions) can be appropriately adjusted by the font size of the characters constituting the watermark, resizing the image, etc. Specific examples of these settings according to the subject will be described later.

[0041] Next, the control unit 11 sets (S42) the color and transmittance of the watermark whose position has been determined (S41) according to the color recognized as described above (S32). Here, the color and transmittance are set so that both the subject and the watermark are easily distinguishable. Here, if there is a distribution of colors in the location where the watermark is added (such as a part of the subject), the color and transmittance of the watermark are not made uniform, and accordingly, the distribution of the color and transmittance is set so that both the subject and the watermark are easily distinguishable. Specific examples of these settings according to the subject will be described later.

[0042] In the operation of Fig. 3, the determination of the marking conditions (S4) was performed for each frame, but in a moving image, the position of the same subject may change within the image. Since the control unit 11 can recognize such a change in the image using image analysis technology, when this is recognized, it is sufficient to recognize only the position of the subject within the image in the image analysis (color recognition of the subject and its surroundings: S32) without performing the video analysis (subject recognition: S31) in the operation of Fig. 4. This makes it unnecessary to perform the operation of Fig. 5 for all frames, thereby shortening the processing time.

[0043] The above operations are performed before the video for distribution is distributed to the user-side playback device 100. FIG. 6 shows the configuration of the user-side playback device (playback means) 100. Here, similar to the video server 10, a control unit 101, an operation unit 102, a display unit 103, a working memory 104, and a storage unit 105 are provided. Here, the user-side playback device 100 is a PC for playing back the video for distribution and displaying it on the display unit 103, and therefore, unlike the video server 10, only data used by itself (including the distributed video for distribution) is stored in the storage unit 105, which has a smaller capacity than the storage unit 15. In addition, a user who uses this can perform authentication processing (such as inputting a password) by operating the operation unit 102. This information is transmitted to the authentication server 50 in FIG. 1, whereby the appropriateness of use (distribution of the video for distribution) is determined.

[0044] This user-side playback device 100 is operated by a user U and can request the video server 10 to distribute video for distribution. At this time, authentication is performed by the authentication server 50, so this distribution is only performed to the registered user U (user-side playback device 100). At the time of this distribution, the encoded video for distribution is sent together with the corresponding marking information. When the control unit 101 decodes this video for distribution and displays it on the display unit 103, it adds a watermark based on this marking information. The watermark displayed here is information that can identify the authenticated user, not the copyright holder.

[0045] This operation will be described below. Fig. 7 is a flow chart showing the operation when distributing such video for distribution. Here, first, a certain user U operates the user-side playback device 100, and in particular requests distribution of a specific video for distribution and logs in (enters authentication conditions) (S51). These authentication conditions are transmitted to the authentication server 50, and the authentication server 50 uses this to determine whether the user is properly registered and distribution is permitted (distribution permitted or not) (S52). If distribution is not permitted (S52: No), the operation ends.

[0046] If distribution is permitted (S52: Yes), the authentication server 50 notifies the video server 10 of this fact, and the video server 10 then causes the sending server 40 to transmit the corresponding distribution video from the recording device 20 in encoded form to the user-side playback device 100, which receives it (S53). Similarly, the video server 10 also causes the user-side playback device 100 to transmit marking information corresponding to the distribution video, which the user-side playback device 100 receives (S54). These are temporarily stored in the storage unit 105 of the user-side playback device 100.

[0047] Next, in the user-side playback device 100, the control unit 101 decodes the received video for distribution, recognizes the image for each frame, and adds a watermark to this image in accordance with the marking conditions (S55). The watermark used here is character information that can identify this user U (e.g., a registered user name), and the control unit 101 can recognize this by an operation for authenticating user U (S51). The control unit 101 can temporarily store the video for distribution after the watermark has been added again in the storage unit 105.

[0048] Thereafter, the control unit 101 causes the display unit 103 to display the distribution video to which the watermark has been added as described above (S56). Note that the operations of S55 and S56 are actually performed in units of GOPs (Group of Pictures). Thereafter, when an operation to stop playback is performed by operating the operation unit 102 of the user-side playback device 100, the control unit 101 ends this playback operation (S56) (S56). At this time, it is preferable that the control unit 101 deletes the distribution video and mark addition information before and after the addition of the watermark that have been temporarily stored in the storage unit 105 as described above.

[0049] By the above operation, only authenticated user U can obtain and view the distribution video on the user-side playback device 100 side. At this time, even if this user U tries to use the displayed video illegally, the video is prevented from being used illegally because a watermark has been added to the video. It is possible to use the screen of the display unit 103 illegally by taking a hard copy of the screen, such as a screenshot, but in this case, the user's identification information is added to the screen as a watermark, so this illegal use is prevented.

[0050] In this case, the content of the watermark displayed will be different for each user U. Furthermore, it is preferable that the video to be distributed before the watermark is added be stored with high confidentiality on the recording device 20 side, which is far from the user U side, and when the user name is used as the watermark as described above, it is preferable that this be added to the video on the side where playback is directly performed (user-side playback device 100) in the configuration of Figure 1. In contrast, when the above configuration is used, it is particularly easy to add a watermark (Figure 7: S55) on the user-side playback device 100 side by using the marking conditions that are preset as described above.

[0051] Note that the content displayed as the watermark does not have to be text information (user name). For example, information other than text information for identifying user U, such as a two-dimensional code corresponding to the identification information of user U, can be used. Also, similar to conventional watermarks, the identification information of the copyright holder of the image and the identification information of user U may be combined and used as the watermark.

[0052] Next, a specific form of adding a watermark to an image (the contents of settings in the mark adding information) will be described. As described above, the watermark used here is displayed overlapping the subject while maintaining the distinctiveness of both the watermark and the subject. FIG. 8 is a simplified example showing such a relationship between the subject and the watermark. Here, one watermark is added to one screen G0. Also, as described above, the video for distribution and the watermark are actually color videos (images), but for the sake of explanation below, they are assumed to be monotone.

[0053] In Fig. 8(a) and (b), the same subject (person) H is present in image G0. As described above, the watermark used here is a user name or the like, but here, for convenience, this watermark M is referred to as "WaterMark". Also, the watermark M is actually semi-transparent, but here, for convenience, it is written in black. Here, in Fig. 8(a), subject H is slightly to the left of the center and appears small (as if it is far away), whereas in Fig. 8(b), subject H is slightly to the right of the center and appears large (as if it is close).

[0054] Here, the watermark M is added so that it partially overlaps with the face, which is the most important part in identifying the subject (person) H. When the images in Figs. 8(a) and (b) are stolen, they are often used as "images of subject H", so by setting it in this way, it is possible to prevent such theft. In this case, by making the watermark M semi-transparent, it is also possible to identify the subject H. For example, in a video (moving image), when the subject H moves in the image from the image in Fig. 8(a) to the image in Fig. 8(b), the watermark M also moves accordingly. Note that the size of the watermark M is the same in Figs. 8(a) and (b), but this size may be changed according to the size of the recognized subject H in the image. In this case, it is preferable to make the watermark M in Fig. 8(b) larger than the watermark M in Fig. 8(a).

[0055] The degree to which the watermark overlaps the subject can be set appropriately to avoid a decrease in the identity of the subject due to the overlap of the watermark. In Fig. 8(a)(b), the watermark M is defined as the subject's line of sight as the important part of the subject, and is arranged so as to completely overlap the line of sight, but if, for example, the identity of the line of sight or aesthetics is important as a characteristic of the image, the degree of overlap can be reduced. If the watermark M is not displayed completely overlapping the important part of the subject, it is preferable to display the watermark M only partially overlapping, or at a position spaced a certain distance away from the important part of the subject, or at a position using image analysis that is close to the important part of the subject and does not impair the identity or aesthetics of the important part of the subject.

[0056] In the example of FIG. 8, only one subject H exists in one screen, but if there are multiple subjects H, a watermark can be added to each subject H in the same manner as described above.

[0057] Also, while the example in Figure 8 is an example of images for each frame (at each time point) of the video to be distributed, in reality, subject H often exists continuously between temporally consecutive frames. In such cases, it is not necessary to display a watermark in all frames in which subject H exists, and it is also possible to provide some frames without a watermark (providing frames without a watermark between frames with a watermark). In this case, the watermark will flash within the video. Even in such cases, the watermark can still deter theft, and the visibility of subject H can be increased.

[0058] Furthermore, even if a user attempts to perform rendering processing following the watermark so that only the watermark portion cannot be identified in order to illegally use the image, the watermark may blink and disappear temporarily, making it difficult to follow using image processing. Even in frames in which the watermark blinks and disappears, there are multiple watermarks displayed, and as long as all of the watermarks do not disappear, the deterrent effect of the watermark against unauthorized use can be maintained. The image for each frame in the video to be distributed may be analyzed, and the object most frequently detected in the center of the image may be estimated to be the subject.

[0059] Also, multiple watermarks can be added in different modes to a single subject H. FIG. 9 shows an example in which two watermarks are added to a single subject H in an image similar to that in FIG. 8(b). Here, a large watermark M1 and a small watermark M2 are used. Both watermarks M1 and M2 overlap with the subject H. Although the watermark M1 is made large, as described above, since the watermark M1 is composed of characters, there are actually many gaps. For this reason, when only the watermark M in FIG. 8 or the watermark M1 in FIG. 9 is used, it may be possible to, for example, enlarge the image and trim only the desired portion for theft.

[0060] In response to this, it is possible to prevent such unauthorized use by using a small watermark M2 in addition, as shown in Fig. 9. In this case, it is preferable that the watermark M2 overlaps a particularly important part of the subject H, such as the eyes of the subject (person) H.

[0061] FIG. 10 shows an example in which the display mode is changed when the size of the watermark is the same. Watermark M3 in FIG. 10(a) is an example in which the characters are normally displayed semi-transparently, and watermark M4 in FIG. 10(b) is an example in which bordered characters are used. In the semi-transparent case shown in FIG. 10(a), the transparency (transmittance) can be set. This transmittance can be set so that both watermark M3 and subject H can be distinguished according to the color of the overlapping subject H. In this case, the transmittance does not need to be uniform in the watermark, and the transmittance can be set to have position dependency according to the color distribution of the overlapping subject H and its surroundings. In this way, the control unit 11 can specify the transmittance for each position in the marking information.

[0062] As described above, the above describes an example in which the video to be distributed and the watermark are monotone, but in reality, they are often color videos (images). In this case, for example, the color of the watermark can be set according to the color of the subject H and its surroundings. For example, if the relationship between the color of the watermark and the color of the original image is determined in advance, such as making the color of the watermark a complementary color to that of the original image in a portion where the watermark and the original image overlap, the control unit 11 (video server 10) can easily perform this setting. Therefore, even if the video to be distributed is a color image, the control unit 11 can easily set the marking information.

[0063] As another aspect of the present invention, assuming that the subject moves in real time, a watermark may always be synthesized on the subject (at the center of the subject) and the video may be output for use in transmission or distribution. Furthermore, it is preferable to synthesize the watermark on the main part of the subject, such as the face (at the center of the main part). This requires some ingenuity in superimposing the watermark on the subject while paying attention to the subject's movement. This is also possible when there are multiple subjects in an image, and for example, image processing may be performed using AI technology to detect pre-registered subjects and synthesize a watermark.

[0064] In addition, in the configuration of FIG. 1, the video server 10, the transmission server 40, and the authentication server 50 are used, but these do not need to be provided separately, and for example, the video server may have all of these functions. As described above, the video to be distributed is created by editing the material data by the editing device 30. At this time, marking information can be created at the same time, and in this case, the editing device can be used as marking information creating means. In addition, the material data is created by, for example, a cameraman using an imaging device, and at this time, information about the subject may be provided by attaching it to this material data. In this case, this information can be used in estimating the subject (S317, S318 in FIG. 5).

[0065] The actual configuration of these is arbitrary as long as it can realize the same functions as those described above. Alternatively, any computer that can access the recording means can be used as the marking information creating means.

[0066] The present invention has been described above based on the embodiment. This embodiment is merely an example, and it will be understood by those skilled in the art that various modifications are possible in the combination of the respective components, and that such modifications are also within the scope of the present invention. [Explanation of symbols]

[0067] 1. Video server system (video editing system) 10 Video server (means for creating marking information) 11, 101 Control unit 12, 102 Operation section 13, 103 Display section 14, 104 Working memory 15, 105 Storage section 16, 106 Interface section 20 Recording device (recording means) 30 Editing Equipment 40 Sending Server 50 Authentication Server 100, 100A to 100C User side playback device (playback means) G0 Image H Subject (person) M, M1~M4 Watermark N Network U, UA~UC users

Claims

1. A video server system that distributes a video for distribution stored in a recording means to a preregistered user and displays the video for distribution on the user's side in a form in which a watermark is added to the video for distribution, a playback means which is operated by the user, and which acquires the video to be distributed only when the user is authenticated, and which adds the watermark to the video to be distributed and displays it; a marking information creating means for analyzing an image of each frame in the video to be distributed to estimate a subject in the image, and by recognizing a position of the subject in the image, creating marking information that defines a position in the image where the watermark should be added so that the subject and the watermark overlap, and a manner of adding the watermark, and storing the marking information in the recording means in combination with the video to be distributed; Equipped with the watermark includes identification information that identifies the user; The regeneration means includes: a video server system which acquires the marking information when acquiring the video to be distributed, and which adds the watermark to the video to be distributed in accordance with the marking information and displays the video to be distributed;

2. 2. The video server system according to claim 1, wherein said marking information creating means creates said marking information so as to display said watermarks of different sizes at a plurality of different positions on a single screen.

3. A video server system as described in claim 1 or 2, characterized in that the marking information creation means creates the marking information so that frames to which the watermark is not added are placed between frames to which the watermark is added in the video to be distributed.

4. 4. The video server system according to claim 1, wherein said marking information creating means creates said marking information so that said watermark is overlapped with a highly distinctive portion of said subject.

5. 5. The video server system according to claim 1, wherein said marking information creating means creates said marking information so that said watermark overlaps a predetermined portion of said subject.

6. 6. The video server system according to claim 1, wherein said marking information creating means creates said marking information so that said watermark overlaps with the center of said subject.

Images