JP6732668B2 - Extraction device, extraction method, and extraction program - Google Patents

Description

The present invention relates to an extraction device, an extraction method and an extraction program.

In recent years, along with the dramatic spread of the Internet, services via networks have been actively provided. As an example of such a service, there is a so-called matching service that introduces the opposite sex to the user. For example, in such a service, the user browses the profile and face image of the opposite sex provided by the service, and selects a desired opposite sex from the browsed.

Here, as a technique related to a face image, there is known a technique of presenting information corresponding to a face image that the user generally prefers while considering a face element that the user likes.

JP, 2013-232093, A

However, the above-mentioned conventional technique cannot always improve the success rate of matching. Specifically, in the above-mentioned conventional technique, only the face image that the user likes is provided. That is, in a matching service or the like, even if the user recognizes the other party as having a desired appearance, it is not clear whether the recognized other person likes the user. Further, even if the user tries to convey his/her appearance to the other party, it is generally difficult to verbalize his/her appearance, and thus it is difficult to accurately convey his/her appearance to the other person. For this reason, in the above-mentioned conventional technology, it is not always possible to improve the probability that both parties recognize each other as their preference in the matching service.

The present application has been made in view of the above, and an object thereof is to provide an extraction device, an extraction method, and an extraction program capable of improving the success rate of matching.

The extraction device according to the present application generates a feature amount space, which is a space in which the target corresponding to each image is plotted at a predetermined position, based on the physical feature of the target included in the image obtained by capturing the target. And a specifying unit that specifies a predetermined range desired by the user in the feature amount space generated by the generating unit based on a predetermined request regarding the target received from the user and the specifying unit. An extraction unit that extracts, from the objects included in a predetermined range, an object that meets a predetermined condition related to the user.

According to one aspect of the embodiment, it is possible to improve the success rate of matching.

FIG. 1 is a diagram illustrating an example of an extraction process according to the embodiment. FIG. 2 is a diagram illustrating a configuration example of the extraction processing system according to the embodiment. FIG. 3 is a diagram illustrating a configuration example of the extraction device according to the embodiment. FIG. 4 is a diagram illustrating an example of the face image storage unit according to the embodiment. FIG. 5 is a diagram illustrating an example of the feature amount space storage unit according to the embodiment. FIG. 6 is a diagram illustrating an example of the user information storage unit according to the embodiment. FIG. 7 is a diagram for explaining an example of the reception process according to the embodiment. FIG. 8 is a flowchart (1) showing a processing procedure according to the embodiment. FIG. 9 is a flowchart (2) showing the processing procedure according to the embodiment. FIG. 10 is a hardware configuration diagram illustrating an example of a computer that realizes the functions of the extraction device.

Hereinafter, modes (hereinafter, referred to as “embodiments”) for implementing an extraction device, an extraction method, and an extraction program according to the present application will be described in detail with reference to the drawings. Note that the extraction device, the extraction method, and the extraction program according to the present application are not limited by this embodiment. Further, the respective embodiments can be appropriately combined within a range in which the processing content is not inconsistent. Also, in each of the following embodiments, the same parts are designated by the same reference numerals, and duplicated description will be omitted.

[1. Example of extraction processing]
First, an example of the extraction process according to the embodiment will be described with reference to FIG. FIG. 1 is a diagram illustrating an example of an extraction process according to the embodiment. FIG. 1 shows an example in which the extraction device 100 according to the present application performs an extraction process for improving the success rate of matching in a service for performing matching between users.

In the embodiment, as an example of a service for performing matching between users, a dating service that provides an encounter between the opposite sex is given. Specifically, in the service according to the embodiment, a user who uses the service is a user who is registered in advance in the service (in the embodiment, a user who is a target to be matched with the user (for example, a user who uses the service). (A user of the opposite sex) is referred to as a “target”), and a target of his/her preference is selected. For example, in a service, a target is matched when a preset condition (a condition required by a partner to be matched) and information of a user who selects the target (hereinafter referred to as “user information”) are matched. The selected target is introduced to the user as a candidate. For example, when the target introduced to the user receives an offer or the like from the user, it is determined that the user and the target are matched.

Generally, in such a dating service, a user registers his/her user information when using the service. For example, the user information is attribute information such as sex, age, annual income, hobbies, etc. of the user and a profile. Further, the user registers the image of himself (for example, a face image) in the service together with the user information. In the service, the face image of the user is posted together with the registered user information, and the user who uses the service browses.

That is, in the service, the user can search for his/her favorite target while looking at the heterosexual face image registered in the service. However, even if the user can select his/her favorite target, it is not certain that the selected target is an appropriate matching partner. Specifically, it is not clear whether or not the target selected by the user likes the appearance of the user, and there is a possibility that the user will make a one-sided selection. This is because it is generally difficult to verbalize the appearance and it is not registered as clear information such as gender and age. In other words, it is difficult to verbalize and register conditions such as “what kind of appearance you want the other person to have,” and it is also difficult to convey the content accurately even when trying to convey your own appearance to the other person. For this reason, it is difficult for the service side to establish a match in which the matched two are in their favorite appearances. That is, even if the user can search for his/her favorite target, the target does not always prefer his/her appearance, and it is not clear whether the selected user and the selected target can be appropriately matched.

Therefore, the extraction device 100 illustrated in FIG. 1 improves the success rate of matching between the user who uses the service and the target desired by the user by using the extraction process described below. Hereinafter, the extraction processing according to the embodiment will be described along the flow with reference to FIG. 1.

The extraction device 100 illustrated in FIG. 1 is a server device that provides a matching service between users. For example, the extraction device 100 functions as a web server on the network and provides the service according to the embodiment.

The user terminals 10 ₁ and 10 ₂ shown in FIG. 1 are information processing terminals used by users who use services. In the example of FIG. 1, the user terminal 10 ₁ is utilized by the user U01, the user terminal 10 ₂ is used by the user U02.

In the following, when it is not necessary to distinguish the user terminal 10 ₁ and the user terminal 10 _2, etc. are simply referred to as "user terminal 10". Moreover, when it is not necessary to distinguish the user U01, the user U02, and the like, they are simply referred to as “user”. In the following, the user may be read as the user terminal 10. For example, reference to "user U01 transmits the face image" may actually may indicate the status of "user terminal 10 ₁ user U01 is utilized to transmit the facial images".

In FIG. 1, the extraction device 100 first receives registration of a service from each user. In this case, the extraction device 100 receives user information from each user who wants to register for the service. The extraction device 100 also receives a face image from each user.

For example, the user U01 who wants to use the service transmits the face image to the extraction device 100 (step S11). For example, the user U01 captures its face image by using the camera function provided in the user terminal 10 _1, transmits the captured facial image to the extraction device 100. Note that the extraction device 100 accepts registration to the service and transmission of face images not only from the user U01 but also from other users (in the example of FIG. 1, user U11, user U12, user U13, etc.). To do.

Then, the extraction device 100 generates a feature amount space based on the face image acquired from the user U01 or the like (step S12). Specifically, the extraction device 100 generates a feature amount space in which each user is plotted (arranged) such that the closer the face images are, the closer the distance is to each other, with respect to all the acquired face images.

For example, the extraction device 100 performs learning on the acquired face image to extract the feature amount of the image. Specifically, the extraction device 100 extracts the element characterizing the face of the subject included in the image by using the method of deep learning. Furthermore, the extraction device 100 expresses the element that characterizes the face as a vector configured by a predetermined number of dimensions (for example, 100 dimensions). Then, the extraction device 100 generates a feature amount space in which the user is plotted at the position on the space indicated by the vector corresponding to the face image of each user. For such processing, various known methods (for example, DCGAN (Deep Convolutional Generative Adversarial Networks) which is a deep learning model) may be used.

In the example of FIG. 1, the extraction device 100 is assumed to generate a different feature amount space for each gender. For example, the extraction device 100 generates a feature amount space V01 which is a space in which the female users registered in the service are plotted. Note that, in FIG. 1, for visualization, each plotted object is visualized by compressing it in three dimensions by using an example in which it is expressed in a three-dimensional space (for example, the space is PCA (Principal Component Analysis)). Although an example) is shown, the feature amount space V01 may not actually be three-dimensionally configured.

As shown in FIG. 1, in the feature amount space V01, each target is plotted on the feature amount space V01 according to the feature amount of the face image. For example, in the feature amount space V01, the user U01, the user U11 who is another female user, the user U12, and the user U13 are plotted as targets.

In the feature amount space V01, since each target is plotted according to the feature amount extracted in the face image of each target, users having similar faces are plotted in the vicinity. For example, when the features of the face images of the user U01 and the user U11 are similar to each other, the user U01 and the user U11 are plotted in the vicinity in the feature amount space V01. If a predetermined range is formed in the feature amount space V01 by dividing the region (coordinates) on the feature amount space V01 by a predetermined threshold value, the user U01 and the user U11 are in the same range D11. Shall belong. When the feature amount extracted from the face image of the user U12 is separated from the user U01 or the user U11 by a predetermined amount or more, the face of the user U12 is not similar to the face of the user U01 or the user U11. In this case, the user U12 belongs to the range D12 which is a range different from the range D11 to which the user U01 and the user U11 belong. When the feature amount extracted from the face image of the user U13 is separated from the user U01, the user U11, and the user U12 by more than a predetermined amount, the face of the user U13 and the faces of the user U01, the user U11, and the user U12. Is not similar to. In this case, the user U13 belongs to the range D11, which is different from the range D11 to which the user U01 and the user U11 belong, and the range D12 to which the user U12 belongs.

In this way, the extraction device 100 is a “similar user” to each female user registered in the service based on the feature amount space V01 generated based on the face image and the predetermined range in the feature amount space V01. Can be classified as

Subsequently, the extraction device 100 sets, for the user U01 registered in the service, information indicating the appearance that the user U01 likes (hereinafter referred to as "preference information") (step S13). For example, the extraction device 100 presents the face image of the opposite user who has been registered in the service to the user U01, and selects one from the presented face images to set the preference information of the user U01. In this case, the extraction device 100 sets the feature amount (vector) of the selected face image as the user U01's preference information.

Alternatively, the extraction device 100 may set what kind of face (specifically, facial feature) the user U01 prefers by a predetermined process. Although details will be described later, for example, the extraction device 100 presents two types of face images of the opposite user to the user U01 who uses the service, and allows the user U01 to select which face is the favorite face. At this time, the extraction device 100 presents to the user U01 two types of face images from which the feature amount has already been extracted. Then, the extraction apparatus 100 generates the preference information of the user U01 by presenting two types of face images of the opposite user and repeating the process of causing the user U01 to select which face is the favorite face. As an example, the extraction device 100 generates the preference information of the user U01 by calculating the average of each vector indicating the feature amount of the face image selected by the user U01. The preference information generated in this way can be said to be a vector indicating the preference of the user U01 in the feature amount space in which the opposite sex user is plotted. The extraction device 100 sets the generated preference information as one of the information associated with the user U01. Similarly, it is assumed that the extraction device 100 also acquires the preference information of the user U11, the user U12, and the user U13 registered in the service and sets it as the information of each user.

Through the above processing, the extraction device 100 generates the feature amount space V01 intended for the female user registered in the service. Note that the extraction device 100 may perform a process of adding the female user to the feature amount space V01 when the female user is newly registered. That is, the extraction device 100 may appropriately perform the process of updating the target included in the feature amount space V01.

After that, the extraction device 100 receives access from the user U02 who is a male user who intends to use the service. The extraction device 100 acquires user information and a face image from the user U02, and also receives preference information of the user U02. That is, the user U02 sets the preference information of the user U02 itself when using the service (step S14).

For example, the extraction device 100 sets the preference information of the user U02 by using the same method as when setting the preference information of the user U01. Then, the extraction device 100 identifies, in the feature amount space V01, a predetermined range corresponding to the preference information of the user U02. Furthermore, the extraction device 100 identifies the target included in the predetermined range corresponding to the preference information of the user U02. After that, the extraction device 100 extracts a target that likes the face of the user U02 from the specified targets (step S15).

The process of step S15 will be specifically described. First, when the vector indicating the preference information of the user U02 indicates the vicinity of the range D11 in the feature amount space V01, the extraction device 100 targets the user U02 and is a user belonging to the range D11. Or user U11 is specified. Subsequently, the extraction device 100 refers to the user U01 belonging to the range D11 and the preference information set for the user U11.

Then, the extraction device 100 extracts, from the objects that the user U02 likes, the objects that the face of the user U02 likes. For example, the extraction device 100 has preference information within the range to which the user U02 belongs in the feature amount space to which the user U02 belongs (in the example of FIG. 1, the feature amount space to the male user) among the targets belonging to the range D11. The users whose vector is heading are extracted. That is, the extraction device 100 extracts a female user whose vector indicating preference information is heading for the space (predetermined range) in which the user U02 is plotted in the feature amount space intended for the male user.

In the example of FIG. 1, it is assumed that the predetermined range to which the target having the facial features preferred by the user U02 belongs is the range D11, and that the users U01 and U11 belong to the range D11. .. Then, it is assumed that the user U02 corresponds to the target of the face preferred by the user U01 and the user U02 does not correspond to the target of the face preferred by the user U11. In this case, the extraction device 100 extracts the user U01 as a user to be matched with the user U02.

In this way, the extraction device 100 extracts the physical characteristics of the target included in the image of the target (the face image in the example of FIG. 1) so that the target corresponding to each image is located at a predetermined position. A feature space V01 that is a plotted space is generated. Then, the extraction device 100 specifies a predetermined range D11 desired by the user U02 in the generated feature amount space V01 by accepting a predetermined request regarding the target from the user U02 who uses the service. Further, the extraction device 100 extracts, from the targets included in the specified range D11, targets that match a predetermined condition regarding the user U02 (in the example of FIG. 1, the user U02's face is a favorite).

As described above, the extraction device 100 according to the embodiment extracts, in the matching service, the user U01 who has the desired appearance of the user U02 who uses the service and who further prefers the appearance of the user U02. .. That is, the extraction device 100 can accurately match users who have their favorite appearances with each other regarding information that is an important element for the matching service such as face and appearance, but is difficult to verbalize. Specifically, the target extracted as a candidate is not the target unilaterally notified of the preference by the user, but the target extracted based on the preference information of the target itself. Accordingly, the extraction device 100 can extract users who like each other's appearance, and thus can improve the success rate of matching (for example, the presented users are established as a pair). Further, according to the above extraction processing, there may be a plurality of candidates (that is, targets to be displayed as recommendations) extracted for the user U02 who uses the service. Even in this case, the candidates extracted by the extraction device 100 are targets that at least prefer the appearance of the user U02. Therefore, the user U02 can easily approach the presented target. As a result, the extraction device 100 can increase the probability of users contacting each other and the possibility of accepting each other, so that the success rate of matching can be improved. Hereinafter, the configuration and the like of the extraction device 100 that performs such processing and the extraction processing system 1 including the extraction device 100 will be described in detail.

[2. Configuration of extraction processing system]
Next, the configuration of the extraction processing system 1 including the extraction device 100 according to the embodiment will be described with reference to FIG. FIG. 2 is a diagram showing a configuration example of the extraction processing system 1 according to the embodiment. As illustrated in FIG. 2, the extraction processing system 1 according to the embodiment includes a user terminal 10 and an extraction device 100. These various devices are connected via a network N so that they can communicate by wire or wirelessly. The extraction processing system 1 shown in FIG. 2 may include a plurality of user terminals 10.

The user terminal 10 is, for example, a smart phone, a desktop PC (Personal Computer), a notebook PC, a tablet terminal, a mobile phone, a PDA (Personal Digital Assistant), an information processing device such as a wearable device (Wearable Device). Is. The user terminal 10 accesses the extraction device 100 and uses the matching service provided by the extraction device 100 according to the operation by the user.

The extraction device 100 is a server device that executes the extraction process according to the embodiment. The extraction device 100 is, for example, a web server and provides a user with a matching service provided on the Internet.

[3. Configuration of extraction device]
Next, the configuration of the extraction device 100 according to the embodiment will be described with reference to FIG. FIG. 3 is a diagram illustrating a configuration example of the extraction device 100 according to the embodiment. As illustrated in FIG. 3, the extraction device 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The extraction device 100 includes an input unit (for example, a keyboard and a mouse) that receives various operations from an administrator who uses the extraction device 100, and a display unit (for example, a liquid crystal display) for displaying various information. You may have.

(About communication unit 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card) or the like. The communication unit 110 is connected to the network N by wire or wirelessly, and transmits and receives information to and from the user terminal 10 and the extraction device 100 via the network N.

(About storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory (Flash Memory), or a storage device such as a hard disk or an optical disk. The storage unit 120 includes a face image storage unit 121, a feature amount space storage unit 122, and a user information storage unit 123.

(About face image storage unit 121)
The face image storage unit 121 stores information about face images received from users registered in the service. Here, FIG. 4 illustrates an example of the face image storage unit 121 according to the embodiment. FIG. 4 is a diagram illustrating an example of the face image storage unit 121 according to the embodiment. In the example shown in FIG. 4, the face image storage unit 121 has items such as “user ID”, “sex”, “face image ID”, and “feature amount information”.

The “user ID” indicates identification information for identifying the user. In addition, in the embodiment, the identification information and the reference numeral used in the description are common. For example, a user whose user ID is “U01” indicates user U01. “Gender” indicates the gender of the user. “Face image ID” indicates identification information for identifying a face image.

The “feature amount information” indicates information about the feature amount extracted from the face image. Note that in the example shown in FIG. 4, the feature amount information is represented by a concept such as “G01”. However, in reality, the feature amount information includes an element corresponding to the feature extracted from the face image and an element. It is represented by the indicated value or the like. For example, the feature amount information is shown as a vector having elements as dimensions and values indicating the elements as numerical values in each dimension. Specifically, the feature amount information is obtained by dimensioning each element that is a feature of the user's face, such as the shape of the entire eye, the shape and color of the white eye, the ratio of the white eye and the black eye, and normalizing each element. It is a digitized vector. Extraction of such feature amount information may be realized by various known techniques (for example, image recognition technique using deep learning) as described above.

That is, in the example of the data shown in FIG. 4, the gender of the user U01 identified by the user ID “U01” is “female”, and the face image identified by the face image ID “P01” is registered. Also, it is indicated that the feature amount information of the face image is “G01”.

(Regarding the feature space storage unit 122)
The feature amount space storage unit 122 stores information about the feature amount space generated by the extraction device 100. Here, FIG. 5 illustrates an example of the feature amount space storage unit 122 according to the embodiment. FIG. 5 is a diagram illustrating an example of the feature amount space storage unit 122 according to the embodiment. In the example shown in FIG. 5, the feature amount space storage unit 122 has items such as “feature amount space ID”, “attribute”, and “face image ID”.

The “feature amount space ID” indicates identification information for identifying the feature amount space. The “attribute” indicates the attribute of the user (target) who belongs to the feature amount space. The “face image ID” indicates identification information of each face image included in the feature amount space. Although not shown in FIG. 5, not only the face image ID but also the information for identifying the user belonging to the feature amount space and the feature amount information of each face image are stored in the feature amount space storage unit 122. It is assumed to be remembered.

That is, in the example of the data illustrated in FIG. 5, the attribute of the user who belongs to the feature amount space V01 identified by the feature amount space ID “V01” is “female”, and the feature amount space V01 includes, for example, a face image. It indicates that the targets associated with the face images identified by the IDs “P01”, “P11”, “P12”, and “P13” belong.

(About the user information storage unit 123)
The user information storage unit 123 stores user information such as preference information received from users who use the service. Here, FIG. 6 illustrates an example of the user information storage unit 123 according to the embodiment. FIG. 6 is a diagram illustrating an example of the user information storage unit 123 according to the embodiment. In the example illustrated in FIG. 6, the user information storage unit 123 has items such as “user ID”, “belonging range”, “preference information”, “target range”, and “matching user candidate”.

The “user ID” corresponds to the same item shown in FIG. The “belonging range” indicates a range in which the user is plotted in the feature amount space. For example, the range is a space having a predetermined width in the feature amount space. Therefore, users belonging to the same range have faces (appearances) having similar characteristics. For example, as illustrated in FIG. 1, the range U11 in the feature amount space V01 includes the user U01 and the user U11 having similar features in the face image.

“Preference information” indicates preference information set for each user. Note that in the example shown in FIG. 6, the preference information is represented by a concept such as “H01”, but in reality, like the feature amount information, the preference information is an element corresponding to the feature extracted from the face image. , A value indicating an element or the like. That is, the preference information is represented as a vector having elements as dimensions and values indicating the elements as numerical values in each dimension. The preference information is not limited to a unique numerical value, and may be a numerical value having a certain width. That is, since the preference information is the information indicating the preference of the user, the preference information is not information indicating a specific person (information that is uniquely determined), but has a width (range) in which the facial features that the user likes can be recognized. It may be a vector.

The “target range” is a range corresponding to the preference information and indicates a range in the feature amount space of the opposite sex. Note that the extraction device 100 does not need to set and set a constant threshold for the items “target range” and “range to which it belongs”, and moves the thresholds that set each range so that the range is determined according to the processing. You may change it. In addition, when the preference information is uniquely determined, the extraction device 100 flexibly sets the range, for example, by setting a range from one point in the feature amount space indicated by the preference information to a predetermined area as a “target range”. May be set.

“Matching user candidate” indicates a user who is a candidate extracted as a target to be matched with the user. In the example of FIG. 6, the matching user candidate is indicated by the reference numeral common to the user ID. For example, when the extraction device 100 satisfies the condition that a user belongs to a range targeted by the user and the user is included in the range targeted by the user, the extraction device 100 selects the target as a matching user candidate. Extract. In other words, the matching user candidate is a user who is a target to be recommended to the user when the user sets preference information in the service.

That is, in the example of the data illustrated in FIG. 6, the range to which the user U01 identified by the user ID “U01” belongs in the feature amount space is “D11”, the preference information is “H01”, and the preference information is The range (target range) determined based on this is “E12”, and the matching user candidate corresponding to the user U01 is “U21” or the like.

(Regarding the control unit 130)
The control unit 130 is, for example, a controller, and is a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like, which stores various programs (an example of an extraction program) stored in a storage device inside the extraction device 100. (Corresponding) is executed by using the RAM as a work area. The control unit 130 is, for example, a controller, and is realized by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

As illustrated in FIG. 3, the control unit 130 includes an acquisition unit 131, a generation unit 132, a reception unit 133, a specification unit 134, an extraction unit 135, and a transmission unit 136, and information described below. Realize or execute processing functions and actions. Note that the internal configuration of the control unit 130 is not limited to the configuration shown in FIG. 3, and may be another configuration as long as it is a configuration for performing information processing described later. Further, the connection relationship between the processing units included in the control unit 130 is not limited to the connection relationship illustrated in FIG. 3 and may be another connection relationship.

(About the acquisition unit 131)
The acquisition unit 131 acquires various kinds of information. For example, the acquisition unit 131 acquires user information of a user who uses the service. The acquisition unit 131 acquires user attribute information as user information. Specifically, the acquisition unit 131 acquires attribute information such as gender, age, occupation, annual income, place of residence of the user.

The acquisition unit 131 also acquires the image of the user. For example, the acquisition unit 131 acquires a face image with the user's face as a subject. The acquisition unit 131 also acquires user preference information. The acquisition unit 131 may acquire the preference information created based on the information received from the user by the reception unit 133 described below.

The acquisition unit 131 stores each acquired information in the storage unit 120. In addition, the acquisition unit 131 may appropriately acquire the information stored in the storage unit 120.

The acquisition unit 131 may acquire information regarding the reaction to the service. For example, the acquisition unit 131 may acquire the user reaction to the candidates extracted in the service. The acquisition unit 131 acquires at least one of the number or rate at which the target is selected (clicked or touched) from the user, or the number or rate at which the user and the target have been successfully matched, as the reaction of the user. To do. That is, the reaction of the user to the candidate is the reaction shown by the user to the target presented as the candidate. For example, when the target presented as the candidate is clicked, the user approaches the candidate. That is, the subject accepted the user's approach, and the two were established as a pair. Alternatively, the acquisition unit 131 may acquire a reaction from the user that the candidate presented to the user is not the user's preference. For example, the acquisition unit 131 may present a message such as "Is this user your favorite?" in the service, and acquire an answer to the message as the user's reaction.

(Regarding the generation unit 132)
The generation unit 132 generates a feature amount space that is a space in which the target corresponding to each image is plotted at a predetermined position by extracting the physical feature of the target included in the image of the captured target.

In the embodiment, the target corresponds to the user who registers in the service provided by the extraction device 100. Further, in the embodiment, the physical feature corresponds to the feature amount information extracted from the face image of the user. That is, the generation unit 132 extracts the feature amount information from the target face image, and generates the feature amount space in which the target is plotted based on the extracted information.

The generation unit 132 extracts face feature amount information from the face image using, for example, a known image recognition technique. Further, the generation unit 132 generates a feature amount space by using, for example, DCGAN.

Further, the generation unit 132 may generate a different feature amount space for each target attribute information. For example, as illustrated in FIG. 1, the generation unit 132 may generate different feature amount spaces for each gender of the target. The generation unit 132 stores the generated information on the feature amount space in the feature amount space storage unit 122.

(Regarding the reception unit 133)
The reception unit 133 receives various types of information. For example, the reception unit 133 receives a service use request from a user. In this case, the reception unit 133 provides a predetermined user interface screen (for example, an application corresponding to the matching service) and prompts the user to register for the service. Then, the reception unit 133 receives the information registered by the user.

When registering for the service, the reception unit 133 receives from the user the conditions desired for the matching target. For example, the reception unit 133 receives conditions such as the target gender, age, occupation, and annual income.

The reception unit 133 also receives preference information regarding the appearance of the target as one of the conditions desired by the user. For example, the reception unit 133 presents the face image from which the feature amount information has already been extracted to the user, and provides the user terminal 10 with a user interface for asking whether the user likes the face. Then, the receiving unit 133 receives the user's preference information in response to the user's reply.

This point will be described with reference to FIG. 7. FIG. 7 is a diagram for explaining an example of the reception process according to the embodiment. FIG. 7 illustrates a situation in which the user terminal 10 displays the user interface 70 (for example, an application screen) provided by the reception unit 133.

As illustrated in FIG. 7, the accepting unit 133 causes the user terminal 10 to display the user interface 70 for selecting the face images indicated by A and B. For example, the user touches the screen with the finger F10 to select which face A or B is preferred.

At this time, the reception unit 133 receives the facial feature amount that the user likes based on the face image that the user has selected (step S21). After accepting the selection from the user, the accepting unit 133 causes the user interface 70 to display two different faces. Then, the receiving unit 133 receives the selection from the user again. The reception unit 133 repeats this selection a predetermined number of times.

Then, the reception unit 133 synthesizes the feature amounts of the selected face image (that is, the face that the user likes) and sets it as preference information of the user (step S22). For example, the reception unit 133 receives information obtained by averaging the feature amount information of each face image selected by the user as the user's preference information.

As shown in FIG. 7, in the service according to the embodiment, a process of referring to the preference information set by the user and the already-set target preference information is performed and matched with the user interface 71. A face image of the candidate is displayed. Such processing is executed by the identifying unit 134, the extracting unit 135, and the transmitting unit 136, which will be described later.

Further, the screen and processing of the user interface 70 shown in FIG. 7 are examples, and the reception unit 133 may receive the preference information by a different method. For example, the user interface 70 may present three or more types of face images instead of presenting two types of face images A and B as shown in FIG. 7. In this case, the accepting unit 133 may accept selection of a plurality of face images instead of accepting selection of only one type of face image. Further, the user may suspend the selection when the desired face is not displayed on the user interface 70. For example, the user makes a selection on the user interface 70, which indicates that "there is no favorite face in this". In this case, the reception unit 133 may display different face images on the user interface 70 and newly receive a desired face from the user. That is, the user may display a different face image and perform the selection operation again. In this way, the reception unit 133 allows each user to select a favorite face image by various methods, and information acquired along with the selection of the face image (for example, vector information corresponding to each face image). Based on the above, preference information may be set for each user.

(About the specifying unit 134)
The specifying unit 134 specifies a predetermined range desired by the user in the feature amount space generated by the generating unit 132 by receiving a predetermined request regarding the target from the user. In the embodiment, the predetermined request is, for example, a request (request) for using the matching service, and also a request for extracting a target that is a candidate for matching with the matching service. The predetermined request includes information about the appearance that the user likes, that is, preference information.

The identifying unit 134 identifies a predetermined range desired by each target by accepting a predetermined request from each target. That is, the specifying unit 134 specifies not only the user who is going to use the service but also the predetermined range desired by each target based on the preference information of the target already registered in the service and plotted in the feature amount space. .. For example, when the feature amount space is generated for each gender as shown in FIG. 1, the specifying unit 134 determines that the target plotted in one feature amount space is the desired range (the range indicated by the preference information). , Specify a range that belongs to another different feature amount space. That is, the identifying unit 134 identifies a predetermined range desired by the user in the feature amount space of the gender different from the user, of the feature amount space generated by the generating unit 132, by receiving a predetermined request regarding the target from the user. To do.

The specifying unit 134 may dynamically set the range with respect to the predetermined range indicated by the preference information. For example, when the user's reaction to the target (candidate) extracted by the extraction unit 135 is acquired, the identifying unit 134 updates the predetermined range associated with the user based on the user's reaction. Good.

For example, a target that belongs to a predetermined range (for example, a range corresponding to the “target range” in FIG. 6) set based on the user preference information and that is extracted by the extraction unit 135 is matched by the user. Is rejected. In this case, the identifying unit 134 may perform a process of shifting the range set as the target of the user in space, a process of narrowing the range, or the like. That is, the specifying unit 134 may set a predetermined range through a predetermined learning process.

(About the extraction unit 135)
The extraction unit 135 extracts, from the targets included in the predetermined range specified by the specifying unit 134, a target that matches a predetermined condition regarding the user. For example, the extraction unit 135 extracts the target when the target satisfies a condition that the target desires a predetermined range including the user and is included in the predetermined range desired by the user.

For example, when a different feature amount space is generated for each gender of the target, the extraction unit 135 desires a predetermined range including the user in the feature amount space in which the user is plotted, and is different from the user. When the condition that the user is included in the predetermined range desired by the gender in the feature amount space is satisfied, the target is extracted as a matching candidate.

The extraction unit 135 may extract a plurality of objects that meet the conditions, or may extract one object that best meets the conditions.

(About the transmitter 136)
The transmission unit 136 transmits various kinds of information. For example, the transmission unit 136 transmits the information regarding the target extracted by the extraction unit 135 to the user terminal 10. For example, the transmission unit 136 transmits, to the user who is the transmission source of the predetermined request, information regarding the target whose preference information, which is the information indicating the preference of the target, of the targets plotted in the feature space matches the user. ..

Specifically, as shown in the user interface 71 shown in FIG. 7, the transmission unit 136 transmits the target face image extracted as a candidate to the user terminal 10.

[4. Processing procedure]
Next, a procedure of processing by the extraction device 100 according to the embodiment will be described with reference to FIGS. 8 and 9. First, a processing procedure regarding generation of the feature amount space will be described with reference to FIG. FIG. 8 is a flowchart (1) showing a processing procedure according to the embodiment.

As shown in FIG. 8, the extraction device 100 determines whether or not a face image has been acquired from a user registered in the service (step S101). When the face image is not acquired (step S101; No), the extraction device 100 waits until the face image is acquired.

On the other hand, when the face image is acquired (step S101; Yes), the extraction device 100 extracts the feature amount information from the face image (step S102). Then, the extraction device 100 plots an object corresponding to the face image based on the feature amount information extracted from the face image to generate a feature amount space (step S103).

Furthermore, the extraction device 100 sets the target preference information plotted in the feature amount space (step S104). Then, the extraction device 100 stores information on the generated feature amount space in the storage unit 120 (step S105).

Next, a processing procedure regarding the extraction processing using the feature amount space will be described with reference to FIG. FIG. 9 is a flowchart (2) showing the processing procedure according to the embodiment.

As shown in FIG. 9, the extraction device 100 determines whether or not a request to use the service has been received from the user (step S201). When the extraction device 100 has not received the service use request (step S201; No), the extraction device 100 waits until the request is received.

On the other hand, when the request to use the service is received (step S201; Yes), the extraction device 100 receives the preference information from the user (step S202). Then, the extraction device 100 specifies the range desired by the user (the range indicated by the preference information) in the feature amount space (step S203).

Furthermore, the extraction device 100 extracts a target that prefers the user from the targets included in the specified range (step S204). Then, the extraction device 100 transmits the extracted target information (for example, the target face image) to the user (step S205).

In addition, the extraction device 100 thereafter responds to whether the target transmitted to the user is actually selected by the user as a pair, or information that the pair has been established (for example, the target of the offer of the pair from the user as a candidate. Information indicating that the user has accepted) or result information indicating that the user's preference and the extracted target are not appropriate. In addition, the extraction device 100 may perform a predetermined learning process, such as changing the setting of a predetermined range to be specified, in response to the result.

[5. Modification example)
The extraction device 100 described above may be implemented in various different forms other than the above embodiment. Therefore, other embodiments of the extraction apparatus 100 will be described below.

[5-1. Service configuration]
In the above embodiment, the example in which the extraction device 100 provides the matching service has been described. However, the matching service may not be provided by the extraction device 100 but may be provided by a predetermined web server or the like. In this case, the extraction device 100 may execute a function as a back-end server that performs a feature amount space generation process, a candidate target extraction process, and the like.

Further, the extraction device 100 may use the extraction processing according to the embodiment not only for a service that performs matching between opposite sex, but also for other services. For example, the extraction device 100 may use the extraction process according to the embodiment for a service for searching for friends.

[5-2. Image information]
In the above embodiment, an example has been shown in which the target feature amount is extracted based on the face image and the preference information is set. However, the extraction device 100 may extract features of various information, not limited to the face. For example, the extraction device 100 may execute the extraction process described in the above embodiment based on the feature amount information of the image of the whole body including the face.

Further, the extraction device 100 may extract not only the face or the whole body but also the feature amount of the appearance including the clothes of the subject in the image. That is, the extraction device 100 may use any information, not limited to the face, as long as the physical characteristics of the target can be extracted using deep learning or the like.

[5-3. Preference information)
The extraction device 100 does not necessarily need to present the target face image registered in the service when setting the preference information. That is, the extraction device 100 may set the user preference information using any method as long as the type (feature) of the user preference can be acquired.

For example, the extraction device 100 acquires images that can be acquired by crawling on the network N, such as entertainers and celebrities. Then, the extraction device 100 extracts the feature amount information from the acquired image. Then, the extraction device 100 presents the acquired image of the celebrity to the user who uses the service, and asks whether or not the celebrity likes this. Then, the extraction device 100 may set the preference information of the user based on the answer from the user.

[5-4. Feature space)
In the above embodiment, the extraction device 100 has shown an example in which a feature amount space that differs for each gender is generated. However, the extraction device 100 may generate one feature amount space and plot the users who use the service in one feature amount space.

In this case, the extraction device 100 calculates, for example, a distance between a vector indicating the user's preference information and a vector belonging to the target range of the user and indicating the target preference information. The extraction device 100 can determine that the smaller the calculated distance is, the closer the user is to the preference of the target. That is, the extraction device 100 may extract the target that is a matching candidate based on the distance between the vectors.

[5-5. conditions〕
In the above embodiment, the example in which the extraction device 100 extracts the candidate targets according to the matching between the user preference information and the target preference information has been described. Here, the extraction device 100 may extract only the objects that further satisfy the conditions by adding the preference information in the face image. That is, the extraction device 100 may extract a target that matches a condition such as sex, age, occupation, or annual income set as a condition for the target and that matches a condition regarding preference information. ..

[5-6. Extraction]
The extraction device 100 may perform the process of transmitting the target as a recommendation to the user without receiving the preference information from the user who uses the service. That is, the extraction device 100 accepts only the face image from the user and plots the user in the feature amount space. Then, the extraction device 100 skips the process of identifying the range based on the preference information from the user among the targets registered in the service, and extracts the targets that the user likes, from all the targets.

According to such a process, the user can receive at least a target recommendation that favors his/her appearance without setting preference information in the service.

[6. Hardware configuration]
The extraction device 100 and the user terminal 10 according to the above-described embodiments are realized by, for example, a computer 1000 having a configuration shown in FIG. Hereinafter, the extraction device 100 will be described as an example. FIG. 10 is a hardware configuration diagram illustrating an example of a computer 1000 that realizes the functions of the extraction device 100. The computer 1000 has a CPU 1100, a RAM 1200, a ROM 1300, an HDD 1400, a communication interface (I/F) 1500, an input/output interface (I/F) 1600, and a media interface (I/F) 1700.

The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400 to control each unit. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 starts up, a program dependent on the hardware of the computer 1000, and the like.

The HDD 1400 stores programs executed by the CPU 1100, data used by the programs, and the like. The communication interface 1500 receives data from another device via the communication network 500 (corresponding to the network N shown in FIG. 2) and sends the data to the CPU 1100, and the data generated by the CPU 1100 is transmitted via the communication network 500 to another. Send to device.

The CPU 1100 controls output devices such as a display and a printer and input devices such as a keyboard and a mouse via the input/output interface 1600. The CPU 1100 acquires data from the input device via the input/output interface 1600. In addition, the CPU 1100 outputs the data generated via the input/output interface 1600 to the output device.

The media interface 1700 reads a program or data stored in the recording medium 1800 and provides the program or data to the CPU 1100 via the RAM 1200. The CPU 1100 loads the program from the recording medium 1800 onto the RAM 1200 via the media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) and a PD (Phase change rewritable Disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. Etc.

For example, when the computer 1000 functions as the extraction device 100 according to the embodiment, the CPU 1100 of the computer 1000 realizes the function of the control unit 130 by executing the program loaded on the RAM 1200. Further, the HDD 1400 stores the data in the storage unit 120. The CPU 1100 of the computer 1000 reads these programs from the recording medium 1800 and executes them, but as another example, these programs may be acquired from another device via the communication network 500.

[7. Other]
Further, of the processes described in the above embodiment, all or part of the processes described as being automatically performed may be manually performed, or the processes described as manually performed may be performed. All or part of the process can be automatically performed by a known method. In addition, the processing procedures, specific names, information including various data and parameters shown in the above-mentioned documents and drawings can be arbitrarily changed unless otherwise specified. For example, the various information shown in each drawing is not limited to the illustrated information.

Further, each component of each device shown in the drawings is functionally conceptual and does not necessarily have to be physically configured as shown. That is, the specific form of distribution/integration of each device is not limited to that shown in the figure, and all or part of the device may be functionally or physically distributed/arranged in arbitrary units according to various loads and usage conditions. It can be integrated and configured. For example, the acquisition unit 131 and the reception unit 133 illustrated in FIG. 3 may be integrated. Further, for example, the information stored in the storage unit 120 may be stored in a predetermined storage device provided outside via the network N.

Further, in the above-described embodiment, an example in which the extraction device 100 performs, for example, an acquisition process of acquiring an image, a generation process of generating a feature amount space, and an extraction process of extracting a target has been shown. However, the extraction device 100 described above may be separated into an acquisition device that performs an acquisition process, a generation device that performs a generation process, and an extraction device that performs an extraction process. In this case, the acquisition device has at least the acquisition unit 131. The generation device includes at least the generation unit 132. The extraction device has at least an extraction unit 135. The process performed by the extraction device 100 according to the embodiment may be realized by the extraction processing system 1 including each of the acquisition device, the generation device, and the extraction device.

Further, the above-described embodiments and modified examples can be appropriately combined within a range in which the processing content is not inconsistent.

[8. effect〕
As described above, the extraction device 100 according to the embodiment has the generation unit 132, the identification unit 134, and the extraction unit 135. The generation unit 132 generates a feature amount space, which is a space in which the target corresponding to each image is plotted at a predetermined position, based on the physical feature of the target included in the captured image of the target. The specifying unit 134 specifies a predetermined range desired by the user in the feature amount space generated by the generating unit 132 based on a predetermined request regarding the target received from the user. The extraction unit 135 extracts, from the targets included in the predetermined range specified by the specifying unit 134, a target that meets a predetermined condition regarding the user.

As described above, the extraction device 100 according to the embodiment does not simply extract the target included in the range desired by the user, but extracts the target that meets the predetermined condition regarding the user. That is, the extraction device 100 extracts a target that matches not the one-sided direction according to the user's request but the predetermined condition required for the user from the target. Thereby, the extraction device 100 can extract the user and the target satisfying the mutual requirements as the matching candidate, and thus the accuracy of the matching can be improved.

Further, the specifying unit 134 specifies a predetermined range desired by each target in the feature amount space by receiving a predetermined request from each target.

In this way, the extraction device 100 according to the embodiment specifies a desired range as a partner with which each of the targets is matched, based on the request (for example, preference information) received from each of the targets. Accordingly, the extraction device 100 can accurately determine whether or not the user meets the target condition in the feature amount space.

Further, the extraction unit 135 extracts the target when the condition satisfies the condition that the target wants a predetermined range including the user and is included in the predetermined range desired by the user.

As described above, the extraction device 100 according to the embodiment extracts a target that belongs to a range desired by the user and that also desires the user (appearance of the user). That is, since the extraction device 100 can extract users who like each other, the users can be matched with each other with high accuracy.

In addition, the generation unit 132 generates a different feature amount space for each gender of the target. The specifying unit 134 specifies a predetermined range desired by the user in the feature amount space of the gender different from the user, of the feature amount space generated by the generating unit 132. The extraction unit 135 sets a condition that the target is in a predetermined range including the user in the feature amount space in which the user is plotted, and that the target is included in the predetermined range in which the user desires in the feature amount space of a gender different from the user. If it satisfies, the target is extracted.

In this way, the extraction device 100 according to the embodiment may generate a different feature amount space for each gender. Accordingly, the extraction device 100 can improve the success rate of matching in the matching service between opposite sexes.

In addition, the extraction device 100 according to the embodiment further includes an acquisition unit 131 that acquires a user's reaction to the target extracted by the extraction unit 135. The specifying unit 134 updates the predetermined range associated with the user, based on the reaction of the user acquired by the acquisition unit 131.

In this way, the extraction device 100 according to the embodiment acquires the user's reaction to the extracted target (for example, the target presented as a candidate in the matching service), and sets the predetermined range based on the acquired information. You may do it. That is, the extraction device 100 can learn and optimize a predetermined range for identifying a matching partner. Thereby, the extraction device 100 can specify the predetermined range that reflects the actual matching tendency.

In addition, the acquisition unit 131 acquires at least one of the number or rate of the target selected by the user or the number or rate of successful matching between the user and the target, as the response of the user to the target.

In this way, the extraction device 100 according to the embodiment may perform the process related to the identification of the predetermined range based on the number of approaches made by the user to the candidates, the number actually established as a pair, and the like. Thereby, the extraction device 100 can specify the predetermined range that reflects the actual matching tendency.

Further, the generation unit 132 generates a feature amount space based on the features of the subject's face in the image. The identifying unit 134 identifies a predetermined range in the feature amount space that includes a target corresponding to the face the user likes, based on a request indicating the user's preference regarding the face. The extraction unit 135 extracts the target included in the predetermined range specified by the specifying unit 134.

In this way, the extraction device 100 according to the embodiment may perform the process based on the feature of the image, particularly the face. Generally, in a matching service or the like, the face can be an important factor when selecting a partner to be a pair. Therefore, the extraction device 100 can improve the success rate of matching by generating the feature amount space based on the features related to the face and performing the extraction process based on the feature amount space.

Further, the extraction device 100 according to the embodiment is a transmission source of a predetermined request, of the objects plotted in the feature amount space, the information regarding the objects whose preference information, which is information indicating the preference of the objects, matches the user. A transmission unit 136 for transmitting to the user is further provided.

As described above, the extraction device 100 according to the embodiment may transmit, as a recommendation, information regarding an object that the user likes to the user. As a result, the extraction device 100 can promote a positive approach of the user and promote encounters between users who have the appearances that they like each other, and thus the number of matchings and the success rate of matching can be improved. ..

The embodiments of the present application have been described in detail above with reference to the drawings. However, this is merely an example, and various modifications and improvements are made based on the knowledge of those skilled in the art, including the modes described in the section of the disclosure of the invention. The present invention can be implemented in other forms described above.

Further, the "section (module, unit)" described above can be read as "means" or "circuit". For example, the acquisition unit can be read as an acquisition unit or an acquisition circuit.

1 Extraction Processing System 10 User Terminal 100 Extraction Device 110 Communication Unit 120 Storage Unit 121 Face Image Storage Unit 122 Feature Space Storage Unit 123 User Information Storage Unit 130 Control Unit 131 Acquisition Unit 132 Generation Unit 133 Reception Unit 134 Identification Unit 135 Extraction Unit 136 transmitter

Claims (10)

A generation unit that generates a feature amount space, which is a space in which the target corresponding to each image is plotted at a predetermined position, based on the physical characteristics of the target included in the image obtained by capturing the target,
Based on a predetermined request regarding the target received from the user, a predetermined range desired by the user is specified in the feature amount space generated by the generation unit, and a physical feature of the user is specified from each of the targets. By receiving a predetermined request regarding the specific amount of each target in the feature amount space ,
An extracting unit that extracts a target that matches a predetermined condition related to the user from the targets included in the predetermined range specified by the specifying unit ;
The generator is
Generate different feature space of the target,
The specific unit is
Of the feature amount space generated by the generation unit, a predetermined range desired by the user is specified in the feature amount space of a gender different from the user,
The extraction unit is
A condition that the target desires a predetermined range including the user in the feature amount space in which the user is plotted, and is included in the predetermined range desired by the user in the feature amount space of a gender different from that of the user, If it satisfies, extract the target,
An extraction device characterized by the above. The extraction unit is
When the target desires a predetermined range including the user and satisfies the condition that the target is included in the predetermined range desired by the user, the target is extracted.
The extraction device according to claim 1 , wherein: A generation unit that generates a feature amount space, which is a space in which the target corresponding to each image is plotted at a predetermined position, based on the physical characteristics of the target included in the image obtained by capturing the target,
A specifying unit that specifies a predetermined range desired by the user in the feature amount space generated by the generating unit based on a predetermined request regarding the target received from the user;
An extraction unit that extracts, from the targets included in the predetermined range specified by the specifying unit, a target that matches a predetermined condition regarding the user,
An acquisition unit that acquires the reaction of the user with respect to the target extracted by the extraction unit,
The specific unit is
Updating the predetermined range associated with the user based on the reaction of the user acquired by the acquisition unit,
An extraction device characterized by the above. The acquisition unit is
As the response of the user to the target, at least one of the number or rate of the target selected from the user, or the number or rate of successful matching between the user and the target is acquired,
The extraction device according to claim 3 , wherein: The generator is
Generating the feature amount space based on the feature of the subject's face in the image,
The specific unit is
Based on a request indicating the user's preference regarding a face, a predetermined range including an object corresponding to the face preferred by the user in the feature amount space is specified,
The extraction unit is
Extracting the objects included in the predetermined range specified by the specifying unit,
The extraction device according to any one of claims 1 to 4 , characterized in that. Among the objects plotted in the feature amount space, a transmission unit that transmits information about a target whose preference information that is information indicating the preference of the target matches the user, to the user who is the transmission source of the predetermined request,
Extraction device according to any one of claims 1 to 5 further comprising a. A computer-implemented extraction method,
A generation step of generating a feature amount space, which is a space in which the target corresponding to each image is plotted at a predetermined position, based on the physical characteristics of the target included in the image obtained by capturing the target,
Based on a predetermined request regarding the target received from the user, a predetermined range desired by the user is specified in the feature amount space generated by the generating step, and the physical feature of the user is specified from each of the targets. A specific step of identifying each predetermined range desired by each target in the feature amount space by receiving a predetermined request regarding
An extraction step of extracting an object that matches a predetermined condition related to the user among the objects included in the predetermined range specified by the specifying step ;
The generation step is
Generate different feature space of the target,
The specific step is
Of the feature amount space generated by the generating step, a predetermined range desired by the user is specified in the feature amount space of a gender different from the user,
The extraction step is
A condition that the target desires a predetermined range including the user in the feature amount space in which the user is plotted, and is included in the predetermined range desired by the user in the feature amount space of a gender different from that of the user, If it satisfies, extract the target,
An extraction method characterized by the above. A computer-implemented extraction method,
A generation step of generating a feature amount space, which is a space in which the target corresponding to each image is plotted at a predetermined position, based on the physical characteristics of the target included in the image obtained by capturing the target,
A specifying step of specifying a predetermined range desired by the user in the feature amount space generated by the generating step based on a predetermined request regarding the target received from the user;
An extraction step of extracting, from the objects included in the predetermined range specified by the specifying step, objects that meet predetermined conditions related to the user;
An acquisition step of acquiring the reaction of the user with respect to the target extracted by the extraction step,
The specific step is
Updating the predetermined range associated with the user, based on the reaction of the user acquired by the acquisition step,
An extraction method characterized by the above. A generation procedure for generating a feature amount space, which is a space in which the target corresponding to each image is plotted at a predetermined position, based on the physical characteristics of the target included in the image obtained by capturing the target,
Based on a predetermined request regarding the target received from the user, a predetermined range desired by the user is specified in the feature amount space generated by the generation procedure, and a physical feature of the user is specified from each of the targets. By receiving a predetermined request regarding the specific space for specifying each predetermined range desired by each target in the feature amount space ,
Among the objects included in the predetermined range specified by the specifying procedure, an extraction procedure of extracting an object that matches a predetermined condition regarding the user, and causing the computer to execute the extraction step .
The generation procedure is
Generate different feature space of the target,
The specific procedure is
Of the feature amount space generated by the generation procedure, a predetermined range desired by the user is specified in the feature amount space of a gender different from the user,
The extraction procedure is
A condition that the target desires a predetermined range including the user in the feature amount space in which the user is plotted, and is included in the predetermined range desired by the user in the feature amount space of a gender different from that of the user, If it satisfies, extract the target,
An extraction program characterized by the above. A generation procedure for generating a feature amount space, which is a space in which the target corresponding to each image is plotted at a predetermined position, based on the physical characteristics of the target included in the image obtained by capturing the target,
A specifying procedure for specifying a predetermined range desired by the user in the feature amount space generated by the generating procedure based on a predetermined request regarding the target received from the user;
An extraction procedure of extracting a target that matches a predetermined condition regarding the user from the targets included in the predetermined range specified by the specifying procedure;
An acquisition procedure for acquiring the reaction of the user with respect to the target extracted by the extraction procedure,
The specific procedure is
Updating the predetermined range associated with the user, based on the reaction of the user acquired by the acquisition procedure,
An extraction program characterized by the above.

Images