JP2011017883A - Target specifying system, target specifying method, advertisement output system, and advertisement output method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system for specifying a person most suitable as an advertising target and displaying an advertisement suitable for the person in real time.SOLUTION: An advertisement output system 1 includes an output device 10 for outputting advertisement video, an advertisement distribution device 30 for distributing contents related to the advertisement video to the output device 10, an image pickup device 20 for acquiring images in a region wherein the advertisement video output from the output device 10 can be viewed, and a target specifying device 40 which analyses the image acquired by the image pickup device 20 to detect viewers of the advertisement video and, specifies a person most suitable as an advertising target among the detected viewers. The advertisement distribution device 30 selects, out of a plurality of contents for advertisement, a content suitable for the person specified by the target specifying device 40 and distributes the content to the output device 10.

Description

  The present invention relates to a technique for specifying a person who is most suitable as an appeal target.

  2. Description of the Related Art Conventionally, a technique (so-called digital signage) that develops an advertisement by displaying an advertisement video on a display device installed in a place that can be seen by the public is well known.

  With regard to such digital signage, for example, Patent Document 1 discloses a technique (advertisement distribution system) aimed at grasping the effect of an advertisement video displayed on a display device.

  This advertisement distribution system captures the vicinity of the display device with a camera and analyzes the captured image to acquire the number of people, the direction of the line of sight, etc., and based on these, the degree of attention (advertisement) of the displayed advertisement video Effect).

JP 2002-269290 A

  By the way, it can be said that the degree of interest in the advertisement video varies among the observers of the advertisement video (that is, an unspecified number of people watching the display screen of the display device). That is, some people consciously look into the advertisement video, and some people accidentally look at the display screen of the display device while walking in the vicinity of the display device.

  For example, if the person who showed the strongest interest in the displayed advertisement video can be identified and the advertisement video suitable for the person can be displayed in real time, the appeal effect of the advertisement video can be expected to be enhanced. .

  However, the above-mentioned Patent Document 1 does not make any mention of the identification of such a person, and no other prior art discloses such a technical idea.

  The present invention has been made in view of the above circumstances, and an appeal target identification system and an appeal target identification method for identifying a person most suitable as an appeal target, and an advertisement output for displaying an advertisement video suitable for the identified person It is an object to provide a system and an advertisement output method.

The appeal target identification system according to the present invention is:
An imaging device that acquires an image of a region where the advertisement video output from the advertisement output device can be observed;
By analyzing the image acquired by the imaging device and detecting a face image of a person reflected in the image, an observer of the advertisement video is detected, and all the detected observations are performed based on a predetermined condition. And an appeal target specifying means for specifying a person most suitable as an appeal target from among the persons.

Further, the appeal target identification method according to the present invention is:
Obtain an image of the area where you can observe the advertisement video output from the advertisement output device,
By analyzing the acquired image and detecting a face image of a person reflected in the image, an observer of the advertisement video is detected, and based on a predetermined condition, among the detected observers The most suitable person to be appealed is identified from the above.

Moreover, the advertisement output system according to the present invention includes:
An advertisement output device for outputting an advertisement video;
An imaging device for acquiring an image of a region where the advertisement video output from the advertisement output device can be observed;
By analyzing the image acquired by the imaging device and detecting a face image of a person reflected in the image, an observer of the advertisement video is detected, and all the detected observations are performed based on a predetermined condition. Appeal target identification means for identifying the most suitable person to appeal from among the persons,
An advertisement distribution unit that selects content suitable for the person identified by the appeal target identification unit from a plurality of content for advertisement, and distributes the content to the advertisement output device;
The advertisement output device outputs an advertisement video related to the content distributed from the advertisement distribution means.

An advertisement output method according to the present invention includes:
Obtain an image of the area where you can observe the advertisement video output from the advertisement output device,
By analyzing the acquired image and detecting a face image of a person reflected in the image, an observer of the advertisement video is detected, and based on a predetermined condition, among the detected observers To identify the most appropriate person to appeal,
Content suitable for the specified person is selected from a plurality of content for advertisement and distributed to the advertisement output device.

  According to the present invention, it is possible to specify a person most suitable as a appeal target with high accuracy. In addition, by displaying an advertisement video suitable for the specified person, it is possible to increase the appeal effect of the advertisement video.

It is a lineblock diagram of an advertising output system concerning one embodiment of the present invention. It is a figure for demonstrating the imaging range by an imaging device, (a) shows the imaging range of a perpendicular direction with the ground, (b) shows the imaging range of the ground and a horizontal direction. It is a block diagram which shows the structure of an advertisement delivery apparatus. It is a figure for demonstrating the table for content selection. It is a block diagram which shows the structure of a appeal target specific apparatus. It is a figure which shows the data structural example of model DB. (A) is a figure for demonstrating a position cost conversion table, (b) is a figure for demonstrating a gaze cost conversion table, (c) is a figure for demonstrating a stay cost conversion table. . It is a flowchart which shows the procedure of appeal target specific processing. It is a figure for demonstrating an observer information table.

  Hereinafter, an embodiment of an advertisement output system according to the present invention will be described with reference to the drawings.

  As illustrated in FIG. 1, the advertisement output system 1 according to the present embodiment includes an output device 10, an imaging device 20, an advertisement distribution device 30, and a appeal target identification device 40. The imaging device 20 and the appeal target specifying device 40 constitute a appeal target specifying system according to the present embodiment of the present invention.

  The output device 10 is composed of, for example, a relatively large display device such as a plasma display or a liquid crystal display, and an acoustic device such as a speaker. The output device 10 is installed in a place where it can be seen by the public such as a street or public transportation. Yes. The output device 10 receives the multimedia data from the advertisement distribution device 30, displays the advertisement video included in the received multimedia data, emits sound, and provides it to the observer OB.

  The imaging device 20 includes a CCD (Charge Coupled Device) camera, a CMOS (Complementary Metal Oxide Semiconductor) sensor camera, and the like, and is installed in the vicinity of the output device 10. In the present embodiment, as shown in FIGS. 2A and 2B, the imaging device 20 is positioned in front of the output device 10 at the center of the upper surface of the output device 10 (position of height L (cm) from the ground surface). It is installed in such a manner that it looks down on the side, and photographs a predetermined area on the front lower side. The imaging device 20 captures an area where the observer OB can view the advertisement video and audio output from the output device 10. In the present embodiment, the height L is about 195 cm, the depression angle θ1 is about 12 degrees, and the angle of view θ2 is about 60 degrees.

  The imaging device 20 always captures the above-described area, acquires a frame image with a frame period of 1/30 seconds, for example, and supplies the frame image to the appeal target identification device 40.

  The advertisement distribution device 30 is connected to the output device 10 via a predetermined network, selects a specific advertisement according to information from the appeal target specifying device 40, and outputs multimedia data corresponding to the selected advertisement to the output device 10. Supply.

  FIG. 3 is a block diagram illustrating an example of the configuration of the advertisement distribution device 30. As illustrated in FIG. 3, the advertisement distribution device 30 includes a content selection memory 300, a content DB 301, a communication unit 302, an input / output unit 303, and a control unit 304.

  The content selection memory 300 stores a content selection table. As shown in FIG. 4, the content selection table includes an address (URL (Uniform Resource Locator)) indicating the storage position of the content (that is, multimedia data to be distributed), its display time, and the viewing target of the content. It is a data table in which person attributes (in this embodiment, age group and sex) most suitable as a person are associated with each other. Here, “a person most suitable as a content viewing target person” means a person who is most likely to have an advertising appeal related to the content.

  The content DB 301 stores content to be distributed (for example, moving images with audio in MPEG format). Each content stored in the content DB 301 is specified by a URL.

  The communication unit 302 communicates with the output device 10, the appeal target specifying device 40, and the like via a predetermined network such as the Internet.

  The input / output unit 303 includes a keyboard, a mouse, a display device, and the like, inputs various instructions and data to the control unit 304, and displays information output from the control unit 304.

  The control unit 304 includes a processor and operates according to a control program stored in a ROM (Read Only Memory) (not shown). For example, the control unit 304 transmits a message requesting transmission of person attribute information (person attribute information request message) to the appeal target identification device 40 via the communication unit 302 in order to determine the content to be distributed next.

  When the control unit 304 receives the person attribute information sent from the appeal target identification device 40 in response to the person attribute information request message, the control unit 304 uses the person attribute (age group and gender) indicated by the person attribute information as a key. Thus, the URL of the corresponding content is acquired from the content selection table. Then, using the acquired URL, the content to be distributed is read from the content DB 301 and supplied to the output device 10 via the communication unit 302.

  When the appeal target specifying device 40 in FIG. 1 receives the above-described person attribute information request message from the advertisement distribution device 30, the appeal target specifying device 40 starts the appeal target specifying process. When the appeal target specifying process is started, the appeal target specifying device 40 analyzes a captured image of the imaging device 20 in units of frames and is watching a advertisement video displayed on the output device 10 (observer OB). Is detected. And one specific person is elected as a subject of appeal from one or a plurality of detected observers OB based on a predetermined condition. In addition, when the detected observer OB is one person, the observer OB is selected as an appeal target.

  When the appeal target specifying device 40 selects a person to be appealed, the appeal target specifying device 40 generates response information (that is, person attribute information) including the attribute (age group, gender) and transmits the response information to the advertisement distribution device 30.

  FIG. 5 is a block diagram illustrating an example of the configuration of the appeal target specifying device 40. As shown in FIG. 5, the appeal target identification device 40 includes a model DB 400, a frame memory 401, a work memory 402, a cost calculation memory 403, an input / output unit 404, a communication unit 405, and a control unit 406. Is composed of.

  As illustrated in FIG. 6, the model DB 400 stores a correspondence relationship between combinations of various feature amounts obtained by analyzing models (populations) of various face images, and age groups and genders.

  The frame memory 401 sequentially stores each frame image supplied from the imaging device 20.

  The work memory 402 functions as a work area for the control unit 406.

  The cost calculation memory 403 stores various cost conversion tables as shown in FIG. The cost conversion table is a data table for deriving the cost from the result of analyzing the frame image. The cost here is a value that serves as an index when selecting an appeal target.

  The position cost conversion table is a cost conversion table used for obtaining the position cost of the observer OB, and shows the correspondence between the position and the position cost in the frame image (for example, a 640 × 480 (pixel) bitmap image). This is shown (see FIG. 7A).

  As can be seen from FIG. 7A, the position cost becomes the highest when the observer OB is located in the place (area) where the advertisement image displayed on the output device 10 is most easily seen. It gets smaller as you leave. That is, a higher position cost is given to an observer OB located in a place where the advertisement image displayed on the output device 10 is easier to see.

  The control unit 406 analyzes the frame image, obtains the coordinates of the center part of the detected face image of the observer OB, and uses the position cost conversion table for the current time of the observer OB (that is, the frame image). Determine the location cost. The control unit 406 obtains the position cost of the observer OB for each analysis process of each frame image, and obtains the final position cost that serves as an index when identifying the person to be appealed by averaging the cost. .

The gaze cost conversion table shown in FIG. 7B is a cost conversion table used for obtaining the gaze cost of the observer OB, and shows a correspondence relationship between the average movement amount of the observer OB and the gaze cost. For each frame image analysis process, the control unit 406 determines the center coordinates (x _i , y _i ) of the face image of the observer OB in the frame image and the observer OB in the frame image analyzed immediately before. The distance (pixel) between the center coordinates (x _i -1, y _i -1) of the face image is obtained and averaged to obtain the average movement amount of the observer OB (the face image) during the frame period. Ask for. When the average movement amount is obtained, the control unit 406 acquires the gaze cost of the observer OB using the gaze cost conversion table.

  As shown in FIG. 7B, in this gaze cost conversion table, the corresponding gaze cost value increases as the average movement amount value decreases. This is because, in general, a person who is watching an advertisement video can be estimated to have little movement.

  The stay cost conversion table shown in FIG. 7C is a cost conversion table used for obtaining the stay cost of the observer OB, and the number of frame images in which the same observer OB is detected in the appeal target specifying process. Shows the correspondence between the cost and residence cost. As can be seen from the contents of this stay cost conversion table, the corresponding stay cost value increases as the number of detected sheets increases.

  Returning to FIG. 5, the input / output unit 404 includes a keyboard, a mouse, a display device, and the like, and inputs various instructions and data to the control unit 406 and displays information output from the control unit 406.

  The communication unit 405 is connected to the imaging device 20 via a predetermined communication cable and communicates with the advertisement distribution device 30 and the like via a predetermined network such as the Internet.

  The control unit 406 includes a processor and the like, operates according to a control program stored in a ROM (not shown), receives a frame image output from the imaging device 20 via the communication unit 405, and stores the frame image in the frame memory 401. To do.

  In addition, when receiving the person attribute information request message from the advertisement distribution device 30, the control unit 406 executes appeal target identification processing. Hereinafter, the details of the appeal target specifying process will be described along the flowchart of FIG.

  First, the control unit 406 acquires one frame image from the frame memory 401 and develops it in the work memory 402 (step S101).

  Next, the control unit 406 analyzes the acquired frame image, and extracts a face image of a person (that is, an observer OB) who is viewing the advertising video being displayed on the display device 20 (step S102).

  The method for extracting the face image of the observer OB is arbitrary. For example, the image is binarized with a threshold value determined based on the average luminance of the entire frame image, and two black spots (estimated as an eye image) at a certain distance (equivalent to 10 to 18 cm) in the binarized image ), And then, in the original frame image, extract a range of images based on the extracted set of black spots, and match this with a sample of facial images prepared in advance, If they match, the extracted image is recognized as the face image of the observer OB.

  In order to improve the detection accuracy of the observer OB, the direction of the line of sight is obtained from the extracted face image, and when the direction is directed to the display screen of the display device 20, the face image is used as the face image of the observer OB. The specification may be recognized.

  Next, the control unit 406 obtains a center position (x, y coordinates) and a plurality of feature quantities for identifying the face for each face image (step S103).

  The “feature amount” refers to a parameter representing the feature of the face image, a gradient vector representing the density gradient of each pixel in the image, color information (hue, saturation), density, texture feature, depth information of each pixel. Any feature such as a feature of an edge included in the image may be expressed. As this feature value, various known feature values can be used. For example, the distance between both eyes and the apex of the nose can be used.

  When the control unit 406 acquires the center position and the feature amount of the face image of each observer OB, the control unit 406 updates the contents of the observer information table using these pieces of information (step S104). The observer information table is a data table in which information of each identified observer is stored in the work memory 402 temporarily by the control unit 406 during the appeal target specifying process. As shown in FIG. 9, the observer information table includes “No.”, “attribute”, “number of continuous detections”, “total position cost”, “total movement amount”, “face position”, “feature amounts 1 to 3”. m ”and other data items.

  In this observer information table, “continuous detection number” includes the continuous number in which the face image of the same observer OB was detected in the analysis of each frame image, that is, the face image of the same observer OB. A numerical value indicating the number of continuous frame images is stored.

  Stored in the “total position cost” is a numerical value indicating the total position cost of each frame image of the same observer OB in time series.

  The “total movement amount” includes the movement distance of the center position of the face image of the same observer OB obtained for each frame image continuous in time series (more specifically, for each frame image after the second frame). A numerical value indicating the total sum of (pixels) is stored.

  In “face position”, coordinates (x, y) indicating the center position of the face image of the observer OB in the frame image analyzed this time are stored in units of pixels.

  The control unit 406 searches each record in the observer information table and determines whether or not a record that matches the center position and the feature amount of the face image of each observer OB acquired this time is included. It can be said that a human can hardly move during a frame period (for example, 1/30 second). For example, in normal walking, only about 10 cm can move within this time. Therefore, the control unit 406 determines that the faces of the same person are within the moving range of about 10 cm from the previous position and have features that substantially match. On the other hand, if the feature amounts are largely different even if the positions are almost the same, or if the positions are greatly shifted even if the feature amounts are substantially the same, it is determined that the faces are different people.

  If there is a matching record, the control unit 406 updates the content of the corresponding data item in the record. This will be specifically described below.

  The control unit 406 obtains the position cost (current position cost) of the observer OB in the frame image analyzed this time. Further, the control unit 406 determines the distance between the coordinates stored in the “face position” of the record (that is, the center coordinates of the face image in the previous frame image) and the center coordinates of the face image acquired this time ( That is, the current movement amount) is obtained.

  Then, the control unit 406 adds +1 to the stored value of “continuous detection number” of the record, adds the current position cost to the stored value of “total position cost”, and moves the current value to the stored value of “total movement amount”. Add the amount. Then, the center coordinates of the face image acquired this time are overwritten in “face position”.

  As described above, the corresponding record in the observer information table is updated.

  On the other hand, if no matching record exists in the observer information table, a new record for the observer OB is added to the observer information table. Hereinafter, this case will be specifically described.

  The control unit 406 compares the feature amount group obtained this time with the feature amount group stored in the model DB 400 (see FIG. 6), and obtains the age group and gender of the observer OB. And stored in the “attribute” of the new record in the observer information table (see FIG. 9). Further, the control unit 406 stores “1” in “continuous detection number”, stores the current position cost in “total position cost”, and stores “0” in “total movement amount”. Then, the center coordinates of the face image acquired this time are stored in “face position”, and the set of feature amounts acquired this time is stored in “feature amounts 1 to m”.

  A new record is added as described above.

  When the face image of the observer OB that has been detected until the previous time is not detected from the current frame image, the control unit 406 removes the observer OB from the appeal target, so that the corresponding record in the observer information table Is deleted.

  When the contents of the observer information table are updated as described above, the control unit 406 determines that the elapsed time (processing time) from the start of this processing (appeal target specifying processing) is a predetermined time (for example, 5 seconds). Is determined (step S105). If the processing time does not exceed the predetermined time (NO in step S105), the control unit 406 acquires the next frame image from the frame memory 401 (step S101), develops it in the work memory 402, and repeats the above-described processing. Do.

  On the other hand, when the processing time exceeds the predetermined time (YES in step S105), the control unit 406 calculates the cost (final cost) of each observer OB based on the contents of each record in the observer information table. (Step S106). The formula for calculating the final cost is shown below.

Final cost = α x position cost + β x gaze cost + γ x dwell cost α, β, and γ are weight adjustment factors

  In the above formula, the position cost is obtained by dividing the “total position cost” by the “continuous detection number”. Further, the gaze cost is obtained by dividing the “total movement amount” by the “continuous detection number” −1 to obtain an average movement amount, and using this average movement amount and the gaze cost conversion table (see FIG. 7B). Derived. Further, the stay cost is derived by converting the “continuation detection number” as the number of detected frames and using the stay cost conversion table (see FIG. 7C).

  As described above, the control unit 406 calculates the final cost of each observer OB, and selects the observer OB having the highest final cost as the appeal target. Then, the control unit 406 generates person attribute information including the “attribute” of the record corresponding to the observer OB, and transmits it to the advertisement distribution apparatus 30 as a response to the person attribute information request message (step S107).

  Thus, the appeal target specifying process ends.

  Upon receiving the person attribute information from the appeal target specifying device 40, the advertisement distribution device 30 selects content suitable for the person attribute (age group and gender) indicated by the person attribute information, and outputs the content via the communication unit 302. 10 is supplied. Thereby, the advertisement image suitable for the observer OB selected as the appeal target is output from the output device 10 and can be provided to the observer OB in real time.

  As described above, according to the advertisement output system 1 of the present embodiment, the imaging device 20 installed in the vicinity of the output device 10 captures a predetermined area in front of the display screen of the output device 10 and obtains it. By analyzing the obtained frame image, a person (observer OB) who can be estimated to be viewing the advertisement video is detected. Then, one person to be appealed is selected from the detected observers OB. In such selection, various costs acquired using various cost conversion tables shown in FIG. 7 are used as indices.

  In the present embodiment, for each observer OB, a position cost that is an index as to whether or not the advertisement video is easily visible, a gaze cost that is an index as to whether or not the advertisement video is being watched, and a viewing time index. Find the residence cost. Then, based on the various costs obtained, a person to be appealed is selected. In this way, since the appeal target is selected by incorporating the indexes based on a plurality of viewpoints, it becomes possible to specify the person most suitable as the appeal target with high accuracy.

  And since the advertisement video suitable for the appeal target person selected in this way can be displayed in real time, it becomes possible to enhance the appeal effect of the advertisement video.

  In addition, this invention is not limited to the said embodiment, Of course, the various change in the range which does not deviate from the summary of this invention is possible.

  For example, in the above embodiment, the information (person attribute) of the appeal target person specified by the appeal target specifying device 40 is used from the viewpoint of enhancing the appeal effect of the advertisement video to be distributed next. There are various ways of using the information. For example, the identified appeal target person can be said to be the person who showed the strongest interest in the displayed advertisement video, so that information may be used as data when analyzing the advertising effectiveness, and its value is sufficiently It can be said that there is.

  Further, in the case where the person to be appealed is specified mainly for use in the analysis of the advertising effect, the object to be appealed identification system (comprised of the imaging device 20 and the appeal target identification apparatus 40) as in the above embodiment. May be distributed in a single system without being incorporated into the advertisement display system 1. In this case, the advertisement to be analyzed does not have to be an advertisement relating to the video displayed on the display device. For example, the imaging device 20 is installed in the vicinity of an advertising billboard or poster installed in a place where it can be seen by the public, and the appeal target identification device 40 is among a plurality of persons who observe the advertisement indicated by the billboard or the like. From the above, the person to be appealed may be specified.

  Moreover, in the said embodiment, although the appeal target specific process is complete | finished when predetermined time passes after the start, the timing which complete | finishes a process is arbitrary. For example, the processing may be terminated when the number of analyzed frame images reaches a predetermined number. Furthermore, there are various timings for starting the appeal target identification process. For example, the appeal target specifying device 40 may periodically execute the appeal target specifying process, or may start the appeal target specifying process in synchronization with the switching of the advertisement video.

  Moreover, in the said embodiment, although the object person of appeal is selected using three types of costs, a position cost, a gaze cost, and a stay cost, an object person of appeal is selected using one or two of these costs. May be elected.

  Further, the stay cost conversion table may associate the stay time of the observer OB with the stay cost.

  In the above embodiment, when updating the observer information table in the appeal target specifying process (step S104 in the flowchart of FIG. 8), the detection of the face image of the observer OB that has been detected until the previous time is interrupted even once. The corresponding record in the observer information table (see FIG. 9) has been deleted, but there is no limitation on the conditions for excluding the detected observer OB from the appeal target. For example, when a face image of a certain observer OB cannot be detected continuously for a predetermined number of times, a certain grace period may be provided such as deleting a record corresponding to the observer OB.

DESCRIPTION OF SYMBOLS 1 Advertisement output system 10 Output device 20 Imaging device 30 Advertisement distribution apparatus 40 Appeal target specific device 300 Content selection memory 301 Content DB
302 Communication Unit 303 Input / Output Unit 304 Control Unit 400 Model DB
401 Frame memory 402 Work memory 403 Cost calculation memory 404 Input / output unit 405 Communication unit 406 Control unit

Claims (7)

An imaging device that acquires an image of a region where the advertisement video output from the advertisement output device can be observed;
By analyzing the image acquired by the imaging device and detecting a face image of a person reflected in the image, an observer of the advertisement video is detected, and all the detected observations are performed based on a predetermined condition. An appeal target identification means for identifying a person most suitable as an appeal target from among the persons,
A target identification system characterized by that. The appeal target specifying means specifies the person most suitable as the appeal target based on the position of the face image of each observer in the acquired image of the imaging device.
The appeal target identification system according to claim 1, wherein: The appeal target specifying means detects the amount of movement of the position of each observer's face image in the acquired image of the imaging device, and based on the detection result, specifies the person most suitable as the appeal target.
The appeal target identification system according to claim 1 or 2, characterized in that The appeal target specifying means specifies the most suitable person as the appeal target based on the observation time of the advertisement video of each observer.
The appeal target identification system according to any one of claims 1 to 3, wherein Obtain an image of the area where you can observe the advertisement video output from the advertisement output device,
By analyzing the acquired image and detecting a face image of a person reflected in the image, an observer of the advertisement video is detected, and based on a predetermined condition, among the detected observers To identify the most suitable person to appeal for,
The appeal target identification method characterized by this. An advertisement output device for outputting an advertisement video;
An imaging device for acquiring an image of a region where the advertisement video output from the advertisement output device can be observed;
By analyzing the image acquired by the imaging device and detecting a face image of a person reflected in the image, an observer of the advertisement video is detected, and all the detected observations are performed based on a predetermined condition. Appeal target identification means for identifying the most suitable person to appeal from among the persons,
An advertisement distribution unit that selects content suitable for the person identified by the appeal target identification unit from a plurality of content for advertisement, and distributes the content to the advertisement output device;
The advertisement output device outputs an advertisement video related to the content distributed from the advertisement distribution means;
An advertisement output system characterized by that. Obtain an image of the area where you can observe the advertisement video output from the advertisement output device,
By analyzing the acquired image and detecting a face image of a person reflected in the image, an observer of the advertisement video is detected, and based on a predetermined condition, among the detected observers To identify the most appropriate person to appeal,
Selecting content suitable for the specified person from among a plurality of contents for advertisement, and delivering to the advertisement output device;
An advertisement output method characterized by the above.

Images