JP5483543B2 - Video distribution service realization apparatus based on viewer identification, video distribution service realization method, video distribution service realization program, and recording medium recording the program - Google Patents

Description

  The present invention relates to a video service realization method based on viewer identification and an apparatus therefor.

Conventional video services can be broadly divided into the following two types.
(1) Representing video for a large number of general viewers without specifying viewers, as represented by television broadcasting.
(2) Video that is requested by the viewer when the viewer takes some action, such as video-on-demand or Youtube.

  Conventionally, as a video-on-demand server device, for example, a device described in Patent Document 1 has been proposed. Moreover, as a technique utilized by this invention and a technique relevant to this invention, there exist a thing of patent document 2, and nonpatent literature 1-7.

JP-A-8-181969 JP 2009-157767 A

Shingo Ando, Yoshinori Kusachi, Akira Suzuki, Kenichi Arakawa, “Method of Estimating Pose of 3D Object Using Support Vector Regression”, IEICE Transactions D, August 2006, Vol. J89-D No. 8, pp. 1840-1847 Shingo Ando, Akira Suzuki, Yuko Takahashi, Takayuki Yasuno, “Fast Object Detection and Identification Using Simultaneous Probability Incremental Code Correlation”, “Image Recognition and Understanding Symposium (MIRU2007)”, July 2007 Yuji Mita, Toshimitsu Kaneko, Osamu Hori, “Probabilistic incremental code correlation suitable for image matching of objects with individual differences”, IEICE Transactions D-II, August 2005, Vol. J88-D-II No. 8, pp. 1614-1623 Fumihiko Saito, “Image matching robust against block matching voting”, IEICE Transactions D-II, October 2001, Vol. J84-D-II No. 10, pp. 2270-2279 Shigeru Akamatsu, “Face Recognition by Computer-Survey”, IEICE Transactions D-II, August 1997, Vol. J80-D-II No. 8, pp. 2031-2046 Yoshiko Sugaya, Shindo Ando, Akira Suzuki, Hideki Koike, “3D FACE RECONSTRUCTION FROM FROM ROTATIONAL MOTION”, IWAIT2009. Erina Ninagawa, Kiyoshi Hosoi, “Automatic gender and age estimation using facial images”, OMRON TECHNICS, 2003, Vol. 43 No. 1 (Vol.145), pp. 37-41

  The above video service has the following problems.

  The method (1) does not necessarily broadcast a video that meets the needs of the viewer because it does not broadcast a video for a specific individual. In addition, when a video for advertising purposes (advertising video) is broadcast, for example, an advertising video of a product targeting an elderly person is broadcast to a younger viewer. The advertisement is not always broadcast, and there is a possibility that an ineffective or less effective advertisement video is broadcast.

  In the method (2), the viewer can acquire a video that meets his needs, but the action for requesting the desired video is a troublesome operation for the viewer, and the user knows the operation method. If not, it will not be possible to obtain video that meets your needs. In addition, since the advertisement video is often not requested and acquired by the viewer, the method (2) is not suitable for broadcasting the advertisement video.

  The conventional method described above has a problem that it is not possible to broadcast a video that meets the viewer's needs or an advertising video that targets the viewer without bothering the viewer.

  SUMMARY OF THE INVENTION The present invention solves the above-mentioned problems, and its purpose is to enable a viewer who can distribute a video meeting the viewer's needs based on the personal information of the identified viewer without performing a cumbersome operation. An object of the present invention is to provide a video delivery service realization apparatus, method, program, and recording medium based on identification.

In order to solve the above problems, an apparatus for realizing a video distribution service based on viewer identification according to the present invention registers personal information and a face image of a viewer in advance, and captures the registered face image of the viewer and an imaging means. The face image recognition means for checking the similarity to the inputted face image of the viewer and identifying the face and outputting the viewer identification result information; and the viewer identification result information identified by the face image recognition means. Based on the function of transmitting to the video distribution means, the video reproduction means having the function of displaying the video data transmitted from the video distribution means on the video display means, and the viewer identification result information transmitted from the video reproduction means Te, determines the image to be broadcast, comprising: a video distribution means for distributing the video data the determined to the video reproduction means, wherein the video distribution device, based on the viewer identification result information, sex and age Ranaru the attribute determination determining viewer attribute of the viewer using a table attribute determination unit attributes defining the multiple layers, depending on the viewer attribute determination unit attribute is determined by, or the viewer identification result information A broadcast list creation unit for creating a list of videos to be broadcast , and a video distribution unit for delivering video data of the list created by the broadcast list creation unit to the video playback means, the broadcast list creation unit Includes a list in which viewer attributes are associated with distribution-scheduled videos, and a broadcast list that creates a broadcast list by extracting distribution-scheduled videos in the list corresponding to the attributes determined by the viewer attribute determination unit a creation unit, the viewer when the attribute determination unit attributes determined by there are multiple, characterized in that to create a broadcast list of delivery order according to the priority of relationship attributes There.

  According to the present invention, for example, a viewer can be identified using an image acquired by a camera installed in the vicinity of a video display device, and a video can be distributed based on the personal information of the identified viewer. This makes it possible to broadcast videos that meet the needs of viewers, and in particular, when distributing videos for advertising purposes, it is possible to distribute advertising videos that match attributes such as the age and gender of viewers. The effect can be enhanced. Also, once the viewer has registered the face image, no action is required to request the video after that, and the face only needs to be directed to a video display device such as a TV near the camera. This eliminates the need for cumbersome operations.

The block diagram which shows the whole embodiment of the video delivery service implementation apparatus based on the viewer identification of this invention. 1 is a block diagram illustrating a configuration of an image processing apparatus according to an embodiment of the present invention. Explanatory drawing which shows the example of a display of the window for registration of the video display apparatus in the example of embodiment of this invention. The flowchart of the face image registration process in the embodiment of this invention. The flowchart of the face identification process in the embodiment of this invention. The block diagram which shows the structure of the video delivery apparatus in the embodiment of this invention. The flowchart of the video delivery process in the embodiment of this invention. Explanatory drawing which shows Example 1 of the image | video control process in the embodiment of this invention. Explanatory drawing which shows Example 2 of the image | video control process in the embodiment of this invention. Explanatory drawing which shows Example 3 of the image | video control process in the embodiment of this invention. Explanatory drawing which shows Example 4 of the image | video control process in the embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings, but the present invention is not limited to the following embodiments.

  FIG. 1 is a block diagram showing an embodiment of a video service realization apparatus based on viewer identification according to the present invention. The viewer watches the video toward the video display device 1 (video display means) such as a television monitor. One or a plurality of cameras 2 (imaging means) are installed in the vicinity of the video display device 1 to photograph the face of the viewer. The camera 2 is connected to an image processing device 3 (face image recognition means) having a face image recognition function by a wireless or wired network. The image processing device 3 uses a camera image to identify a viewer.

  The image processing device 3 resides in or is connected to a video playback device 4 (video playback means) such as a set-top box, and when registering a viewer's face image, a window for registration work in the video display device 1. Is displayed. Also, the viewer identification result (viewer identification result information) output from the image processing device 3 passes through the video reproduction device 4 and the video distribution device 6 (video distribution) of the broadcasting station that distributes the video using the network 5. Means). The video distribution device 6 controls video distribution based on the viewer identification result, and distributes video data corresponding to the viewer to the video reproduction device 4. The video playback device 4 displays the received video data on the video display device 1.

  The video display device 1 is constituted by a display of a computer terminal, for example.

  FIG. 2 is a block diagram illustrating a configuration of the image processing apparatus 3. The image processing apparatus 3 includes a face image registration unit 31 and a face identification unit 32. The face identification unit 32 detects a person's face area from the camera image, and identifies the viewer by collating with a face image registered in advance in the face image registration unit 31 (for example, as described in Patent Document 2). Identify the method). The face identification unit 32 outputs a viewer identification result (user personal information such as user name, gender, date of birth) at each processing timing. When a plurality of people are detected and identified, a plurality of viewer identification results are output. Further, when face detection is performed but identification is not possible, such as when the viewer's face image is not registered, only the face detection count is output.

  The face image registration unit 31 registers personal information input from the viewer and the face image of the viewer imaged by the camera 2 and creates identification data. An image input unit 11, a face detection unit 12, a face posture estimation unit 13, a face re-cutout unit 14, a face shape reconstruction unit 15, and a registered face storage unit 16 are provided.

  The face identification unit 32 inputs a viewer's recognition face image captured by the camera 2, detects an area with a face from the input recognition face image, and detects the detected face area. A face obtained by estimating the posture of the face based on the rotation and rotating the viewer's face image indicated by the identification data created by the face image registration unit 31 to the same posture as the estimated face posture A face is identified by examining the similarity between the image and the inputted recognition face image, and viewer identification result information is output. The recognition image input unit 17 described in Patent Document 2, A face detection unit 18, a face posture estimation unit 19, a face re-cutout unit 20, a registered face rotation unit 21, a face identification unit 22, and a result output unit 23 are provided.

  Non-patent documents 1 to 5 describe techniques related to the face identification unit 32.

  In this embodiment, face recognition is performed using the techniques described in Non-Patent Document 1 and Non-Patent Document 2, but the present invention is not limited to this. In this embodiment, first, the method of Non-Patent Document 1 is applied to detect a human face area from an image and estimate the posture of the face using data for face detection and posture estimation prepared in advance. . Next, the method of Non-Patent Document 2 is applied, and the personal identification of the detected face is performed using the face image of the viewer prepared in advance. At this time, based on the estimated posture, by using the viewer face image of the posture, personal identification with high accuracy is possible.

  The face image registration unit 31 registers in advance the face image of the viewer to be identified. The registration work is performed using a registration window displayed on the video display device 1. FIG. 3 shows an example of a registration window. The personal information display unit 311 inputs the viewer's user name, gender, and date of birth using a character input device such as a keyboard or remote control. In addition to these, it is also possible to input personal information such as age, hobby, place of residence, blood type, occupation, family structure. In the camera image display unit 312 (captured image display unit), the video captured by the camera 2 is displayed in real time, and a guideline such as a contour line or a cross line imitating the shape near the head of a person is superimposed and displayed. Is done. When a viewer registers his / her face image, these guidelines are used as auxiliary lines for head alignment.

  For face images, there are a method for registering only the front direction, and a method for registering faces at various angles in the vertical and horizontal directions in addition to the front direction, depending on the face identification processing method. In the present embodiment, a method for performing face identification using Non-Patent Document 1 and Non-Patent Document 2 described above will be described. Faces in the vertical and horizontal directions are registered in addition to the front direction. There are a plurality of methods for registering face images. The simplest method is (1) Checking the camera image display unit 312 and manually capturing and registering face images facing each other one by one. It is a way to go. As another method, (2) using the face detection function and the face posture estimation function used by the face identification unit 32, the viewer moves the face up and down, left and right in front of the camera, thereby changing the face area of the viewer. There is a method of automatically detecting and estimating the posture and registering it as a face image in the corresponding direction. In this embodiment, the latter method (2) will be described. However, the present invention is not limited to this.

  The registered face image display unit 313 displays a plurality of registered face images, and the viewer can check whether the registered face image is appropriate as a face image at a predetermined angle. If it is determined to be inappropriate, the face image can be deleted and re-registered. In the example of FIG. 3, the number of face images to be registered is 15 in total, 5 in the horizontal direction and 3 in the vertical direction, but any number of faces such as 21 in the horizontal direction and 3 in the vertical direction is 21 in total. Images can be registered. When face images in all directions are not registered, the missing face image is used as a face image in the vicinity. In addition, the registered face image display unit 313 exemplifies face images in a direction desired to be registered using computer graphics, photographs, and the like, so that the viewer can know what direction the face image should be registered. It has been devised.

  As another method for registering a plurality of face images, (3) a method of acquiring three-dimensional shape information of a face portion, reconstructing and registering face images of a plurality of postures therefrom, is also conceivable. The acquisition of the three-dimensional shape information includes a method using a technique described in Non-Patent Document 6 in addition to a method using a three-dimensional digitizer using a laser or a camera.

  Note that the flowchart of the process performed by the face image registration unit 31 is as shown in FIG. First, in step S11, the face image registration unit 31 inputs personal information input from the viewer using a character input device such as a keyboard or a remote controller.

  Next, the face image registration process described in FIG. 2 is performed until an appropriate image is registered (steps S12 and S13).

  In step S14, identification data is created based on the inputted personal information and the registered face image.

  Next, in step S15, it is determined whether or not to register another viewer. If YES, the process returns to step S11, and if NO, the process ends.

  Further, a flowchart of processing performed by the face identifying unit 32 is as shown in FIG. First, in step S21, a recognition image captured by the camera 2 is input.

  Next, in step S22, using the face detection data, a region with a face is detected from the input recognition image by applying the method described in Non-Patent Document 3, for example.

  Next, in step S23, using the face posture estimation data, for example, a face posture is estimated for the detected face region by applying the method described in Non-Patent Document 1, for example.

  Next, in step S24, using the identification data created by the face image registration unit 31, the method described in, for example, Patent Literature 2 and Non-Patent Literature 2 is applied, based on the estimated posture. The viewer identification of the detected face is performed, and the viewer identification result is output.

  The video playback device 4 is a device for displaying video data transmitted from a broadcasting station on the video display device 1, such as a cable TV set-top box or an IP TV tuner, and has a bidirectional communication function. The viewer identification result output from the image processing device 3 is first input to the video reproduction device 4 and transmitted to the video distribution device 6 existing on the broadcasting station side via the network 5. The transmission timing is arbitrary and can be set in consideration of the traffic of the network 5 and the video distribution device 6.

  The video distribution device 6 is a device for distributing video data existing in a broadcast station that distributes video, and has a bidirectional communication function. FIG. 6 shows a configuration example of the video distribution device 6. The video control unit 61 includes a viewer identification result input unit 611 that receives the viewer identification result transmitted from the video reproduction device 4, a viewer attribute determination unit 612 that determines attributes such as sex and age, and the determination result. A broadcast list creation unit 613 that creates a broadcast list for the viewer. The video distribution unit 62 distributes the video data in the video database 621 to the video reproduction device 4 based on the broadcast list created by the video control unit 61.

  A flowchart of processing performed in the video distribution device 6 is as shown in FIG. First, in step S <b> 31, the viewer attribute determination unit 612 determines the viewer attribute for the viewer identification result input to the viewer identification result input unit 611.

  Next, in step S32, the broadcast list creation unit 613 creates a broadcast list for the viewer from the attribute determination result.

  In step S33, the video distribution unit 62 distributes video data to the video reproduction device 4 based on the created broadcast list.

  8 to 11 show examples of processing performed by the video control unit 61. FIG.

  FIG. 8 is an example in which “user name: electric father” is identified. If the personal information “sex: male, date of birth: January 1, 1970” is given to “electric father,” the viewer identification result input unit 611 has “user name: electric father, sex”. : Male, date of birth: January 1, 1970 "is input. The viewer attribute determination unit 612 determines the viewer attribute from the gender and date of birth of the viewer identification result.

  As an example of the viewer attribute, there are eight-layer attributes consisting of gender and age used in the marketing field, etc. Specifically, the C-layer (4-12 years old), the T-layer (13-19 years old) , M1 layer (male 20-34 years old), M2 layer (male 35-49 years old), M3 layer (male 50 years old ~), F1 layer (female 20-34 years old), F2 layer (female 35-49 years old), It is F3 layer (female 50 years old ~). In addition, a method for determining viewer attributes such as a method for determining viewer attributes only by gender, a method for determining age in increments of 10 years, and a method for determining age by roughly classifying into three levels is arbitrary. It is.

  In the example of FIG. 8, since it is a man and 39 years old, it determines with a viewer attribute being a M2 layer. The broadcast list creation unit 613 creates a video distribution list based on the determined viewer attributes. Here, a case where a commercial video is distributed is shown as an example. It is assumed that ten commercial videos scheduled to be distributed are stored in the broadcasting station, and target commercial attributes are set for each commercial. An example is shown in the CM list 6131. In the CM list 6131, there are two commercial videos for the M2 layer (CM_M2 (1) and CM_M2 (2)), and since the viewer attribute is the M2 layer, these two commercial videos are included in the CM list 6131. Candidates for video to be distributed to the viewer. Among these candidate videos, a method of distributing all of them, a method of distributing a randomly selected one, a method of determining the number and order of distribution in advance and distributing according to the rules, etc. are conceivable.

  The example of FIG. 8 shows a method of distributing everything, and CM_M2 (1) and CM_M2 (2) are set in the broadcast list 6132. There is a method of setting the airing order at random or a method of setting in a predetermined order. Here, the airing is performed in a predetermined order (the number in the parenthesis of the CM name indicates the priority of the airing order). An example of setting a broadcast list is shown.

  In the example of FIG. 9, two viewers are identified, and the viewer attributes are discriminated as the M2 layer and the C layer as in the example described with reference to FIG. In this case, four commercial videos (CM_C (1), CM_C (2), CM_M2 (1), CM_M2 (2)) corresponding to the attributes of two viewers are candidates for video to be distributed to the viewers. Become. When video candidates span a plurality of attributes, there are a method of randomly setting the broadcast order and a method of setting according to a predetermined rule. Here, when determining the broadcast order, F1 layer> F2 layer> F3 layer It is assumed that the priority relationship of attributes is used:> M1 layer> M2 layer> M3 layer> T layer> C layer.

  That is, in the example of FIG. 9, the video for the M2 layer has a higher priority than the video for the C layer, and the video for the M2 layer is set first in the order in the broadcast list. Next, the broadcast order of a plurality of commercial videos assigned to each attribute is determined. Here, as in the example of FIG. 8, a predetermined rule (the number in parentheses of the CM name is the priority of the broadcast order). The broadcast list 6132 is set to CM_M2 (1), CM_M2 (2), CM_C (1), and CM_C (2).

  In the example of FIG. 10, two viewers are detected, but the user name is “unknown” due to reasons such as unregistered face images or failure in identification processing. In such a case, the process in the viewer attribute determination unit 612 is skipped, and the process proceeds to the broadcast list creation unit 613. In this case, all videos existing in the CM list 6131 are distribution candidates. As for what is distributed in what order from among these videos, a method of randomly distributing, a method of randomly distributing one by one for each attribute, a method of distributing in accordance with a predetermined rule, etc. can be considered. However, in accordance with the relationship representing the priority of the viewer attribute described in the example of FIG. 9, all the commercial videos existing in the CM list are set in the broadcast list.

  In the example of FIG. 11, two viewers are detected, but one of them is “unknown”. The identified attribute of “electric mother” is determined as the F1 layer, and attribute determination of the “unknown” viewer is skipped. In the broadcast list creation unit 613, CM_F1 (1), which is a commercial video targeting the F1 layer, is set in the broadcast list 6132.

  In the above-described embodiment, when the viewer cannot be identified, the processing in the viewer attribute determination unit 612 is skipped, but the age and gender are estimated from the face image, and the viewer attribute is based on this. There is also a method of judging. There is a technique described in Non-Patent Document 7 regarding a method for estimating age sex from a face image.

  Further, in the processing example in the video control unit 61 described above, the video to be distributed is controlled using only the viewer's attributes, but the control method using the viewer's personal information other than the age gender, Embodiments such as a method of controlling using the number of users, a method of controlling using continuous viewing time, and a method of controlling using time and time zone are also conceivable.

  As for the timing of creating the broadcast list 6132, a method that is performed every time a viewer identification result is input, a method that is performed after all the videos set in the already created broadcast list are broadcast, and the like are conceivable. .

  In addition, the video distribution device 6 can measure the number of viewers and the audience rating by counting the viewer identification results transmitted from a plurality of homes via the network 5. In this case, the number of viewers and the audience rating can be aggregated by attribute using viewer personal information. The video distribution device 6 may have a function of displaying the viewer identification result, the number of viewers, and the audience rating in a graph format or a table format.

  In addition, some or all of the functions of each means in the video distribution service realization apparatus based on viewer identification according to the present embodiment are configured by a computer program, and the program is executed using the computer to realize the present invention. Needless to say, the procedure in the video delivery service realization method based on the viewer identification according to the present embodiment can be configured by a computer program and the program can be executed by the computer. The program is stored in a computer-readable recording medium such as FD (Floppy (registered trademark) Disk), MO (Magneto-Optical disk), ROM (Read Only Memory), memory card, CD (Compact Disc). k) -ROM, DVD (Digital Versatile Disk) -ROM, CD-R, CD-RW, HDD, removable disk, etc., and can be stored and distributed. It is also possible to provide the above program through a network such as the Internet or electronic mail.

DESCRIPTION OF SYMBOLS 1 ... Video display apparatus 2 ... Camera 3 ... Image processing apparatus 4 ... Video reproduction apparatus 5 ... Network 6 ... Video delivery apparatus 11 ... Registration image input part 12, 18 ... Face detection part 13, 19 ... Face posture estimation part 14, DESCRIPTION OF SYMBOLS 20 ... Face re-cutting part 15 ... Face shape reconstruction part 16 ... Registered face memory | storage part 17 ... Image input part 21 for recognition 21 ... Registered face rotation part 22 ... Face identification part 23 ... Result output part 61 ... Image | video control part 62 ... Image | video Distribution unit 311 ... Personal information display unit 312 ... Camera image display unit 313 ... Registered face image display unit 611 ... Viewer identification result input unit 612 ... Viewer attribute determination unit 613 ... Broadcast list creation unit 621 ... Video database 6131 ... CM list 6132 ... Broadcast list

Claims (6)

The viewer's personal information and face image are registered in advance, the face is identified by checking the similarity between the registered viewer's face image and the viewer's face image captured and input by the imaging means, and viewing Face image recognition means for outputting person identification result information;
A video reproduction means having a function of transmitting the viewer identification result information identified by the face image recognition means to a video distribution means, and a function of displaying the video data transmitted from the video distribution means on the video display means;
Video distribution means for determining a video to be broadcast based on the viewer identification result information transmitted from the video reproduction means and distributing the determined video data to the video reproduction means ,
The video distribution means includes
A viewer attribute determination unit that determines an attribute of the viewer using an attribute determination table that defines attributes of a plurality of layers including gender and age based on the viewer identification result information;
A broadcast list creation unit that creates a list of videos to be broadcast according to the attribute determined by the viewer attribute determination unit or the viewer identification result information;
A video distribution unit that distributes the video data of the list created by the broadcast list creation unit to the video reproduction means ,
The broadcast list creation unit includes a list in which viewer attributes and distribution schedule videos are associated with each other, and extracts the broadcast schedule videos in the list corresponding to the attributes determined by the viewer attribute determination unit. A broadcast list creation unit that creates a broadcast list in a distribution order according to a priority relationship of attributes when there are a plurality of attributes determined by the viewer attribute determination unit. An apparatus for realizing a video distribution service based on viewer identification. The video display means includes a personal information display section that displays personal information input from a viewer, a captured image display section that displays a face image of the viewer imaged by the imaging means, and a face image recognition means. A registered face image display unit for displaying the face image of the viewer to be registered,
The face image recognition means includes
A face image registration unit for registering personal information input from the viewer and the face image of the viewer imaged by the imaging means, and creating identification data;
The viewer's face image for recognition of the viewer imaged by the imaging means is input, an area with a face is detected from the input face image for recognition, and the posture of the face based on the detected face area The face image of the viewer indicated by the identification data created by the face image registration unit is rotated so as to have the same posture as the estimated face posture, and the input identifying the face by examining the similarity between recognition of the face image, based on the viewer identification according to claim 1, characterized in that and a face identification unit that outputs the identification result information of the viewer Video distribution service realization device. The face image recognition means registers the viewer's personal information and face image in advance, and checks the similarity between the registered viewer's face image and the viewer's face image captured and input by the imaging means. A face image recognition step of identifying a face and outputting viewer identification result information;
A step of transmitting the viewer identification result information identified by the face image recognition unit to the video distribution unit;
A video distribution step in which video distribution means determines a video to be broadcast based on the viewer identification result information transmitted from the video reproduction means, and distributes the determined video data to the video reproduction means;
Video reproduction means comprising displaying the video data transmitted from the video distribution means on the video display means ,
The video distribution step includes:
A viewer attribute determination step of determining a viewer attribute using an attribute determination table defining a plurality of attributes of gender and age based on the viewer identification result information transmitted from the video reproduction means;
A broadcast list creation step of creating a list of videos to be broadcast according to the attribute determined in the viewer attribute determination step or the viewer identification result information;
A video distribution step of distributing the video data of the list created by the broadcast list creation step to the video reproduction means ,
The broadcast list creating step is a step of creating a broadcast list by extracting a distribution schedule video in a list corresponding to the attribute of the viewer and the distribution schedule video corresponding to the attribute determined in the viewer attribute determination step. When there are a plurality of attributes determined by the viewer attribute determination step, a broadcast list having a distribution order according to the attribute priority relationship is created. Delivery service realization method. The face image recognition step includes
The face image registration unit of the face image recognition unit displays the personal information input from the viewer on the video display unit, and displays the viewer's face image captured by the imaging unit on the video display unit. A step of displaying the face image of the viewer registered in the face image recognition means on the video display means, and registering the personal information input from the viewer and the face image of the viewer imaged by the imaging means A face image registration step for creating identification data;
The face identification unit of the face image recognition means inputs a face image for viewer recognition taken by the image pickup means, detects a region with a face from the input face image for recognition, and detects the detection The posture of the face is estimated based on the determined face region, and the face image of the viewer indicated by the identification data created in the face image registration step is set to the same posture as the estimated face posture. A face identifying step of identifying a face by examining a similarity between the rotated face image and the input recognition face image, and outputting viewer identification result information. The video delivery service realization method based on the viewer identification according to claim 3 . A program for realizing a video distribution service based on viewer identification that causes a computer to function as each means according to claim 1 . A computer-readable recording medium on which is recorded a video delivery service realization program based on viewer identification that causes a computer to function as each means according to claim 1 .

Images