Control device, control method, and program

Abstract

The present invention provides a control device, a control method, and a program for playing avatar content on a content playback device that can improve the recognition of an avatar. [Solution] The control device of this disclosure comprises a person detection unit that detects a person associated with a content playback device, and a playback control unit that, at the timing when the person is detected, causes the content playback device to play avatar content including an avatar that performs a predetermined action on the person.

Description

Technical Field

[0006] , , ,

[0005] , ,

[0007] , ,

[0001] The present disclosure relates to a control device, a control method, and a program for controlling a content playback device.

Background Art

[0002] AI (Artificial Intelligence) avatars that can communicate with humans have been developed using technologies such as AI. Patent Document 1 discloses an avatar register system that displays an avatar on digital signage in a store and allows the avatar to communicate with customers.

Prior Art Documents

Patent Documents

[0003] <0000​​​​​​​​​​​​​​​​​​​​​​​​​​​A control method according to one aspect of the present disclosure is a control method performed by a computer, wherein the computer performs a process of detecting a related person associated with a content playback device, and, at the timing when the related person is detected, causes the content playback device to play avatar content including an avatar that performs a predetermined action on the related person.

[0008] A program according to one aspect of this disclosure causes a computer to perform a procedure for detecting a person associated with a content playback device, and a procedure for causing the content playback device to play avatar content, including an avatar that performs a predetermined action on the person associated with the person, at the time the person associated is detected. [Effects of the Invention]

[0009] According to this disclosure, it is possible to increase the recognition of avatars among a large number of people. [Brief explanation of the drawing]

[0010] [Figure 1] This figure shows an example configuration of a content playback system including a control device according to an embodiment of the present disclosure. [Figure 2] Block diagram showing an example of the hardware configuration of the control unit. [Figure 3] Block diagram showing an example of the hardware configuration of a content playback device. [Figure 4] A diagram illustrating an example of the functional configuration of a control device. [Figure 5] Diagram showing an example of a content playback device screen display. [Figure 6] A flowchart illustrating an example of the overall operation flow of a control device. [Modes for carrying out the invention]

[0011] Embodiments of this disclosure will be described below with reference to the drawings. However, this disclosure is not limited to the examples described in the embodiments. In the following description, components having the same function and configuration will be denoted by the same reference numerals, and their descriptions will be omitted.

[0012] <System Configuration> Figure 1 is a diagram showing an example configuration of a content playback system including a control device according to an embodiment of the present disclosure.

[0013] In the example shown in Figure 1, the content playback system 100 includes a control device 10, a content playback device 20, a generation model 30, and a network 40.

[0014] The control device 10 is a device that controls the playback of content in the content playback device 20. The control device 10 is a type of computer, such as a PC (Personal Computer), smartphone, or tablet device.

[0015] The content playback device 20 plays various types of content in accordance with the control of the control device 10.

[0016] The content playback device 20 is, for example, a digital signage. In this specification, the digital signage, as the content playback device 20, is assumed to be installed in a corner of a passageway or plaza in various facilities such as stores, train stations, and government offices.

[0017] The control device 10 controls the playback of content in the content playback device 20. The control device 10 causes the content playback device 20 to play predetermined content and avatar content, as described below. The predetermined content is content that includes images. Images include moving images or still images (including slideshows of still images). The predetermined content may also include audio associated with the images. In this disclosure, the content of the predetermined content is not particularly limited, but for example, advertising content such as commercials and product introductions is envisioned.

[0018] Avatar content is content that includes an avatar that performs a predetermined action for a person related to the content playback device 20.

[0019] In the present specification, an avatar is a virtual character, and avatar content is content that includes how the avatar performs various actions (motions). Specifically, avatar content is composed of image content (preferably, moving image content) that shows the movement of the avatar. Further, the avatar content may include audio related to the movement of the avatar. The audio in the avatar content may include, in addition to the voice emitted by the avatar, sound effects that match the movement of the avatar.

[0020] A person related to the content playback device 20 is, for example, a person present in front of the display screen of the digital signage as the content playback device 20. As described above, since the content playback device 20 is installed in passages, squares, etc. of various facilities, the person related is a person who passes in front of the display screen of the content playback device 20. The control device 10 causes the content playback device 20 to play back the avatar content at the timing when the person related is detected. The timing when the person related is detected is specifically the timing when the person related passes in front of the display screen of the content playback device 20.

[0021] The control device 10 causes the generation model 30 to generate the content of the predetermined action performed by the avatar in the avatar content.

[0022] Generative model 30 is a deep learning model that includes a vast number of parameters. An example of generative model 30 is a large language model (LLMS) that is generated by deep learning based on a large amount of text data and is useful in natural language processing. A large language model can interpret the content of input text and generate an appropriate response according to the content. An example of a large language model is GPT (Generative Pre-trained Transformer) developed by OpenAI.

[0023] In this disclosure, the generative model 30 is not limited to a large-scale language model. Furthermore, the data input to the generative model 30 is not limited to text format data. The generative model 30 may be a model that takes data other than text format as input and outputs data other than text format as a response. Examples of data other than text format include image format, audio format, or a format that combines text, images, and audio.

[0024] Network 40 is a public network, such as the internet.

[0025] The control device 10 and the content playback device 20 are connected via a network 40 so as to be able to communicate with each other. The content playback device 20 and the generation model 30 are also connected via the network 40 so as to be able to communicate with each other. In the example shown in Figure 1, the network through which the control device 10 and the content playback device 20 communicate, and the network through which the content playback device 20 and the generation model 30 communicate, are the same network 40; however, this disclosure is not limited to this example. The network through which the control device 10 and the content playback device 20 communicate, and the network through which the content playback device 20 and the generation model 30 communicate, may be different networks.

[0026] The control device 10, upon detecting a relevant person, causes the content playback device 20 to play avatar content. The avatar content includes first avatar content in which the avatar performs a first action to attract the attention of the relevant person, and second avatar content in which the avatar performs a second action to speak to the relevant person. The control device 10, upon detecting a relevant person, causes the content playback device 20 to start playing the first avatar content. This makes it possible to pique the interest of a relevant person who happens to pass near the display screen of the content playback device 20.

[0027] Furthermore, while the control device 10 is playing the first avatar content, it causes the generation model 30 to generate the content of the second action in which the avatar speaks to the relevant person in the second avatar content. Then, after playing the first avatar content, the control device 10 plays the second avatar content, which was created based on the content of the second action generated by the generation model 30, to the content playback device 20.

[0028] The content of the second action in the second avatar content, that is, what the avatar says to the related person, is generated by the generative model 30 based on the related person's physical characteristics (e.g., apparent gender, age group, hairstyle, clothing, makeup, etc.) based on an image of the related person. Therefore, avatar content can be played that allows the avatar to make an appropriate conversation with a related person who has shown interest in the avatar.

[0029] The control device 10 plays the second avatar content immediately after the first avatar content has finished playing. For this reason, it is preferable that the first avatar content and the second avatar content be configured to be a single, continuous piece of content.

[0030] <Hardware configuration of control device 10> Figure 2 is a block diagram showing an example of the hardware configuration of the control device 10. The control device 10 is a computer comprising a processor 11, a communication interface (I / F) 12, and memory 13.

[0031] The processor 11 is, for example, a CPU (Central Processing Unit). The processor 11 realizes various functions of the control device 10 by reading and executing programs stored in the memory 13.

[0032] The communication interface 12 connects the control device 10 to the content playback device 20 and the generation model 30 via the network 40, enabling communication between them.

[0033] Memory 13 is composed of, for example, ROM (Read Only Memory), RAM (Random Access Memory), and either an HDD (Hard Disk Drive) or an SSD (Solid State Drive), or a combination thereof. Memory 13 stores programs that the processor 11 reads and executes, as well as various data used for program execution.

[0034] <Configuration of content playback device 20> Figure 3 is a block diagram showing an example of the configuration of the content playback device 20. The content playback device 20 includes a communication I / F 21, an image output unit 22, an audio output unit 23, and an imaging unit 24.

[0035] The communication interface 21 connects the content playback device 20 to the control device 10 via the network 40, enabling communication between them.

[0036] The image output unit 22 is a display device such as a liquid crystal display or an organic EL (Electro-Luminescence) display. The image output unit 22 displays images included in the content when the content playback device 20 plays the content based on the control device 10. The image output unit 22 has a display screen.

[0037] The audio output unit 23 is, for example, an audio playback device such as a speaker. The audio output unit 23 plays the audio contained in the content when the content playback device 20 plays the content based on the control of the control device 10.

[0038] The imaging unit 24 is a camera device that captures a relevant area within a predetermined distance from the display screen of the image output unit 22 and generates a region image. The region image includes at least a relevant person who is within the relevant area. The predetermined distance defining the relevant area can be set in advance by the administrator of the content playback system 100, for example, based on the size of the display screen of the image output unit 22 or the environment of the installation location of the display screen (size, brightness, number of people, etc.). The imaging unit 24 generates a region image, either a moving image or a still image captured at predetermined time intervals.

[0039] <Functional configuration of the control device 10> Figure 4 is a diagram illustrating an example of the functional configuration of the control device 10. In the example shown in Figure 4, the control device 10 comprises a communication unit 111, an image acquisition unit 112, a person detection unit 113, a playback control unit 114, an image extraction unit 115, a generation control unit 116, and a storage unit 117. These functional configurations of the control device 10 are realized by the processor 11 shown in Figure 2 reading and executing a program stored in the memory 13.

[0040] The communication unit 111 communicates with other components of the content playback system 100 (content playback device 20, generation model 30) via the communication interface 12.

[0041] The image acquisition unit 112 acquires a region image from the imaging unit 24 of the content playback device 20.

[0042] The person detection unit 113 detects people (related people) within a region based on the region image acquired by the image acquisition unit 112.

[0043] The playback control unit 114 causes the content playback device 20 to play pre-set content (such as commercials or product introduction videos, as described above) while the person detection unit 113 has not detected any relevant persons. The pre-set content can be stored in the storage unit 117 beforehand, for example.

[0044] Then, when the person detection unit 113 detects a relevant person, the playback control unit 114 plays avatar content, including an avatar that performs a predetermined action on the relevant person, on the content playback device. In this disclosure, the avatar content is content that is different from the predetermined content. The avatar's face, body shape, clothing, etc. in the avatar content may be set in advance by the administrator of the content playback system 100 or the like.

[0045] As described above, the avatar content includes first avatar content, which includes an avatar that performs a first action to attract the attention of the relevant person, and second avatar content, which includes an avatar that performs a second action to speak to the relevant person.

[0046] The playback control unit 114 starts playing the first avatar content when the person detection unit 113 detects a relevant person. After playing the first avatar content, it then plays the second avatar content. This allows the avatar to sufficiently attract the attention of the relevant person before speaking to them.

[0047] Examples of the first actions an avatar might perform in the first avatar content include raising a hand, beckoning, saying a predetermined phrase (such as "Hello!"), walking or running around the screen, falling and getting up, moving towards the foreground of the screen (becoming larger), and leaning out of the screen.

[0048] The playback control unit 114 can attract the attention of relevant individuals by having the content playback device 20 play back the first avatar content, which includes the avatar performing such a first action.

[0049] Figure 5 shows examples of the display screen when predetermined content is being played on the content playback device 20, and the display screen when the first avatar content is displayed as an interruption when a related person is detected. The upper part of Figure 5 shows an example of the display screen when predetermined content is being played, and the lower part of Figure 5 shows an example of the display screen when the first avatar content is displayed as an interruption.

[0050] In the example shown in Figure 5, the avatar is represented by a pictogram-style image, but in reality, the avatar may consist of a realistic or anime-style character. Furthermore, the avatar may be a 2D image or a 3D model.

[0051] In the example shown in Figure 5, the avatar content, including the avatar, is simply displayed in front of the designated content on the screen. However, the avatar content could also be made more prominent by, for example, displaying the designated content in a darker color than the avatar content.

[0052] The first avatar content is created in advance and stored in the memory unit 117. The content of the first action in the first avatar content may be set in advance by the administrator of the content playback system 100 or the like. Multiple first avatar contents, each in which the avatar performs a different first action, may be prepared in advance. In this case, the playback control unit 114 may appropriately (for example, randomly) select one of the multiple first avatar contents and have the content playback device 20 play it.

[0053] When the playback control unit 114 has finished playing the first avatar content, it instructs the content playback device 20 to start playing the second avatar content. As described above, in the second avatar content, the avatar performs a second action in which it speaks to the relevant person.

[0054] The second avatar content is content generated by the generation control unit 115 based on a second action generated by the generation control unit 116 in the generation model 30 while the first avatar content is being played back on the content playback device 20. For this reason, it is preferable that the playback time of the first avatar content be considered to be sufficient for the generation control unit 116 to generate the second avatar content. The method for generating the second avatar content will be described later.

[0055] When the playback control unit 114 has finished playing the avatar content, including the first avatar content and the second avatar content, it causes the content playback device 20 to play predetermined content.

[0056] Furthermore, if a related person is detected while the content playback device 20 is playing predetermined content, the playback control unit 114 will play the avatar content in the foreground and the predetermined content in the background. Specifically, the playback control unit 114 controls the content playback device 20 so that the actions of the avatar in the avatar content appear to be taking place in front of the predetermined content on the screen.

[0057] This makes it possible to give relevant individuals who are within a predetermined distance from the display screen of the content playback device 20 the impression that the playback of avatar content has started in a way that interrupts the predetermined content. This makes it possible to effectively pique the interest of relevant individuals in the avatar content.

[0058] Furthermore, while the content playback device 20 is playing avatar content, the playback control unit 114 may continue playing a predetermined content in the background or pause it. If it pauses, the playback control unit 114 may resume playback of the predetermined content after the playback of the avatar content is completed. Also, if the predetermined content includes audio, the playback control unit 114 may stop playing the audio of the predetermined content or reduce the volume and play it in the background while the avatar content is being played.

[0059] The image extraction unit 115 performs image analysis on the region image acquired by the image acquisition unit 112, detects relevant persons present within the region, extracts them from the region image, and generates an extracted image. If multiple relevant persons are present within the region image, the image acquisition unit 112 can perform extraction for each relevant person.

[0060] The generation control unit 116 causes the generation model 30 to generate the content of the second action in the second avatar content, in which the avatar speaks to the relevant person. The content of the second action includes at least the content of the speech to be uttered by the avatar, and may further include actions (gestures) that the avatar performs in accordance with the speech.

[0061] The generation control unit 116 generates the image extracted by the image extraction unit 115 and a prompt (instruction statement) for generating the second action, and transmits it to the generation model 30 via the communication unit 111. This allows the generation model 30 to generate the content of the second action.

[0062] The prompt content should be such that it appropriately generates what the avatar says to the relevant person based on the attached cropped image, instructing the generative model 30 to do. For example, the prompt could be something like the following:

[0063] The attached image shows your own perspective. You are at the company booth of [Company Name]. First, focus on a person within your field of vision and capture their clothing, belongings, and other characteristics in detail. Remember to capture them as a passerby; do not try to identify them personally. Next, create a message to call out to the target person. The message should include the characteristics you've captured so that the person being addressed can recognize that they are being called. The message should be around 200 characters long and be humorous enough to get the target person to turn around and talk to you. If the person in question is facing you or looking directly at the camera, assume they are interested in you and include something that might interest them in your message, such as, "Are you interested? Will you come closer?" When addressing someone, please use phrases such as "you (referring to a specific person)" or "customer (referring to a specific customer)." (Example) "Hey you with the nice T-shirt, you look like you're having fun. I have something I'd like to show you! Please come closer. That drink looks good, I'm getting thirsty too... why don't you come over for a bit? That's a nice phone case, are you taking a picture? I'd be happy if you could take a nice one! Come on, come even closer!"

[0064] Thus, the prompt may be plain text and may be a list of elements to be instructed to the generative model 30. The prompt may be, for example, a bulleted list. The present disclosure does not particularly limit the format of the prompt and it may be in any format.

[0065] The elements that the generation control unit 116 should include in the prompt to instruct the generation model 30 can be pre-configured by the administrator of the content playback system 100 or the like.

[0066] The prompt exemplified above causes the generative model 30 to generate only the call-and-response text. However, the prompt may also cause the generative model 30 to generate gestures (such as body language) that the avatar will use when calling out to the relevant person, in addition to the call-and-response text.

[0067] The generation control unit 116 transmits the generated prompt and the extracted image to the generation model 30 via the communication unit 111. This allows the generation model 30 to generate the content of the second action, i.e., the content of the call message.

[0068] The generation control unit 116 may transmit to the generation model 30 information regarding the content of predetermined content that was being played in the content playback device 20 before the first avatar content was played, in addition to the prompt and the extracted image. In this case, the content of the second action in the second avatar content can be made to match the content of the predetermined content. Such second avatar content makes it possible to achieve both increased recognition of the avatar among relevant individuals and the advertising effect of the advertising content that the content playback device 20 had previously played.

[0069] When the generation control unit 116 receives the content of the second action generated by the generation model 30 via the communication unit 111, it generates the second avatar content based on that content.

[0070] In this case, the generation control unit 116 can generate the image components of the second avatar content using a known 3D image generation program or the like, with the avatar's face, body shape, clothing, etc., which have been pre-set by the administrator of the content playback system 100, applied to the avatar's 3D data. The generation control unit 116 can also generate the audio components of the second avatar content using a known audio synthesis program or the like, with the avatar's audio synthesis data, etc., which have been created in advance. Finally, the generation control unit 116 can generate the second avatar content by combining the generated image components and audio components.

[0071] Furthermore, it is preferable that the generation control unit 116 generates the second avatar content so that it is continuous with the first avatar content. Specifically, it is preferable to set the position, posture, and movement of the avatar at the start of the second avatar content based on the position, posture, and movement of the avatar at the completion of the first avatar content so that the avatar appears to be moving continuously.

[0072] The memory unit 117 stores various data and programs necessary for the operation of the control device 10. Specifically, the memory unit 117 stores the first avatar content, 3D data of the avatar, data for avatar voice synthesis, a 3D image generation program, a voice synthesis program, and the like.

[0073] With the above configuration, the control device 10 can cause the content playback device 20 to play predetermined content, and when it detects a relevant person, it can attract the attention of the relevant person with the first avatar content, and then have the avatar speak to the relevant person with the second avatar content. By repeatedly executing this series of processes, and having the avatar actively speak to a new person who passes in front of the display screen of the content playback device 20 as a relevant person, the recognition of the avatar can be improved.

[0074] <Example of operation> The following describes an example of the operation of the control device 10. Figure 6 is a flowchart illustrating an example of the operation of the control device 10.

[0075] In step S1, the control device 10 causes the content playback device 20 to play a predetermined content.

[0076] In step S2, the control device 10 acquires a region image of a relevant area within a predetermined distance from the display screen of the content playback device 20.

[0077] In step S3, the control device 10 detects the relevant person within the relevant region based on the region image.

[0078] The processing in step S1 and the processing in steps S2 and S3 are carried out in parallel.

[0079] In step S4, if the control device 10 detects a relevant person within the relevant area in step S3 (step S4:Y), it proceeds to steps S5 and S7. If the relevant person is not detected within the relevant area (step S4:N), it returns to steps S1 and S2.

[0080] In step S5, the control device 10 causes the content playback device 20 to play the first avatar content to attract the attention of the relevant person. At this time, the control device 10 controls the content playback device 20 to play the first avatar content in the foreground and predetermined content in the background.

[0081] In step S6, the control device 10 determines whether or not the playback of the first avatar content has been completed. If the control device 10 determines that the playback of the first avatar content has been completed (step S6:Y), it proceeds to step S11. If it determines that the playback of the first avatar content has not been completed (step S6:N), it repeats the process in step S5 until playback is completed.

[0082] In step S7, the control device 10 performs image analysis on the region image acquired in step S2, extracts the relevant person, and generates a cropped image.

[0083] In step S8, the control device 10 generates a prompt to cause the generation model 30 to generate the content of the second action to be performed by the avatar in the second avatar content, and sends it to the generation model 30 along with the extracted image.

[0084] In step S9, the control device 10 obtains from the generation model 30 the generation result, which includes the content of the second action generated by the generation model 30 based on the prompt and the extracted image.

[0085] In step S10, the control device 10 generates a second avatar content based on the generation results obtained from the generation model 30.

[0086] The processes in steps S5 and S6 and the processes from steps S7 to S10 are carried out in parallel.

[0087] In step S11, the control device 10 causes the content playback device 20 to play the second avatar content generated in step S10.

[0088] When the playback of the second avatar content is completed in step S11, the control device 10 repeatedly executes the series of processes from step S1 to step S11. This allows the avatar to actively speak to each new person who passes in front of the display screen of the content playback device 20, treating that person as a relevant person.

[0089] Furthermore, after the playback of the second avatar content in step S11 is completed, the control device 10 may have the content playback device 20 play a third avatar content that includes a third action in which the avatar engages in conversation with the related person, following the second avatar content. The control device 10 can generate the third avatar content by recording the voice spoken by the related person using a microphone or the like, recognizing the content of the voice using a known voice recognition program, and transmitting the recognized content to a generation model. In this case, the control device 10 can start a new series of processes at the timing when the conversation between the avatar and the related person is completed (when the playback of the third avatar content is completed).

[0090] In this way, when a person passes near the display screen of the content playback device 20 that is playing predetermined content, the control device 10 can play a first avatar content, which includes an avatar that actively calls out to the person, in a manner that interrupts the predetermined content. This makes it possible to pique the interest of people passing in front of the content playback device 20, thereby increasing the recognition of the avatar among a large number of people.

[0091] Furthermore, while the control device 10 is playing the first avatar content on the content playback device 20, the control device 10 can cause the generation model 30 to generate the content of a call message in which the avatar actively addresses the relevant person, and generate a second avatar content in which the avatar addresses the relevant person based on that summary.

[0092] Here, the generative model 30 extracts the characteristics of the relevant person based on an image of the relevant person, and generates the content of the call-and-response message based on those characteristics. This allows the avatar to make a call-and-response message tailored to each individual relevant person. As a result, a relevant person who has become interested in the avatar through the first avatar content will hear the avatar make a call-and-response message that matches their characteristics in the second avatar content, which is played after the first avatar content. This further deepens the relevant person's interest in the avatar.

[0093] Furthermore, when generating the content of the appeal message using the generation model 30, it is also possible to have the model take into account the content of a predetermined item. In this case, it is possible to achieve both increased recognition of the avatar and the promotional effect that is the purpose of the predetermined item.

[0094] <Variation> The embodiments described above are merely one example of the forms that can be taken when implementing this disclosure, and this disclosure can take various modifications other than those described in the embodiments.

[0095] In the embodiment described above, the imaging unit 24 of the content playback device 20 captures a relevant area within a predetermined distance from the display screen and generates a region image. However, this disclosure is not limited to this. For example, the region image may be captured and generated by an imaging device independent of the content playback device.

[0096] In the embodiment described above, the person detection unit 113 of the control device 10 detects a person present in the region based on the region image acquired by the image acquisition unit 112, but this disclosure is not limited to this. For example, a person in the region may be detected by a motion sensor that detects a person near the display screen of a content playback device.

[0097] In the embodiment described above, a digital signage display is assumed as the content playback device 20, and the relevant person is defined as a person located in a relevant area within a predetermined distance from the display screen of the content playback device 20. This disclosure is not limited thereto, and for example, the content playback device may be a terminal device owned and used by an individual, such as a smartphone, tablet, or PC.

[0098] In this case, for example, the control device could play predetermined content when the owner approaches within a predetermined distance from the terminal device's display screen, and then interrupt and play the first avatar content when the owner looks directly at the screen. [Industrial applicability]

[0099] This disclosure is useful for a control device that interrupts predetermined content with the playback of avatar content. [Explanation of symbols]

[0100] 100 Content Playback Systems 10 Control device 11 processors 111 Communications Department 112 Image acquisition unit 113 Person Detection Unit 114 Regeneration Control Unit 115 Image extraction section 116 Generation Control Unit 117 Storage section 12 Communication I / F 13 memory 20 Content playback device 21 Playback I / F 22 Image output section 23 Audio output section 24 Imaging Department 30 Generative Models 40 Networks

Claims

1. A person detection unit that detects related persons associated with a content playback device, A playback control unit causes the content playback device to play avatar content, including an avatar that performs a predetermined action on the aforementioned related person, at the time the related person is detected. A control device equipped with the following features.

2. If the playback control unit detects the related person while playback of predetermined content different from the avatar content is in progress, it will play the avatar content in the foreground and the predetermined content in the background. The control device according to claim 1.

3. If the playback control unit detects the related person while playback of predetermined content different from the avatar content is in progress, it will play the avatar content in the foreground and pause the predetermined content. The control device according to claim 1.

4. The avatar content includes a first avatar content which includes the avatar that performs a first action to attract the attention of the related person, and a second avatar content which includes the avatar that performs a second action which is to speak to the related person. The control device according to claim 2 or 3.

5. A generation control unit that causes the content of the preceding 2 action to be generated in the generation model, A communication unit that receives the content of the second action generated by the generation model, The control device according to claim 4, further comprising the following:

6. The playback control unit, after playing the first avatar content at the timing, plays the second avatar content based on the content of the second action generated by the generation model. The control device according to claim 5.

7. The system further includes an image acquisition unit that acquires a region image of a relevant area within a predetermined distance from the screen of the content playback device. The generation control unit causes the generation model to generate the content of the second action based on the region image. The control device according to claim 5.

8. The system further comprises an image extraction unit that extracts the relevant person from the aforementioned region image and generates an extracted image, The generation control unit causes the generation model to generate the content of the second action based on the extracted image. The control device according to claim 7.

9. The generation control unit extracts the features of the related person based on the cropped image, and generates the content of the second action in the generation model based on the extracted features. The control device according to claim 8.

10. The generation control unit causes the generation model to generate the content of the second action based on the content of the predetermined content. The control device according to claim 9.

11. The person detection unit detects a person as a related person when the person passes through a related area within a predetermined distance from the screen of the content playback device. The control device according to claim 2 or 3.

12. A control method performed by a computer, The aforementioned computer, A process for detecting related individuals associated with a content playback device, The process involves, at the time the aforementioned related person is detected, causing the content playback device to play avatar content including an avatar that performs a predetermined action on the aforementioned related person, A control method for executing this.

13. On the computer, Procedures for detecting individuals associated with a content playback device, The procedure involves, at the time the aforementioned related person is detected, playing avatar content, including an avatar that performs a predetermined action on the aforementioned related person, on the content playback device, A program that executes something.

Images