JP2012173761A - Information processor, information processing method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technology for acquiring information on the sex and clothes of a person included in an image through a simple method.SOLUTION: The information processor comprises: a storing part that stores templates for shapes of persons and shapes of clothes prepared for each type of clothes; and a control part that collates an edge detected in an image captured by a camera with a template, identifies the edge with an edge of a person included in the image, acquires the edge identified with the edge of the person included in the image, for each person, as a shape of the person, estimates the sex of the person related to the shape of the person based on the shape characteristics which is acquired from the shape of the person and enables estimation of the sex, estimates the type of clothes of the person related to the shape of the person by collating the shape of the clothes and the shape of the person, and counts the number of persons in the image for each type of the sex and clothes based on the estimated sex and type of the clothes.

Description

  The present invention relates to a technique for acquiring information related to clothes of a person included in an image captured using a camera.

  Conventionally, various techniques for counting people included in image data captured by a street camera or a portable camera have been proposed.

  For example, Patent Document 1 discloses a technique for counting people by analyzing image data and extracting a person's face image. Japanese Patent Application Laid-Open No. 2004-228561 discloses a technique for extracting the features of a person's face by performing image analysis on image data and totaling the number of persons for each sex, age, and person type.

  In addition, Patent Document 3 discloses a technique for continuously collecting information on a moving monitoring object.

JP 2010-250571 A JP 2007-58828 A JP 2004-274309 A

  However, in the techniques disclosed in these, since the feature of a person existing in an image captured by a street camera or a mobile terminal camera is acquired by face recognition based on image analysis, the processing load is large. In the techniques disclosed in these, it is not possible to acquire and provide information such as clothes and equipment of passersby in a specific place, that is, an area imaged by a street camera or a mobile terminal camera, etc. by a simple method. .

  The present invention has been made in consideration of such points, and is a technique that makes it possible to obtain information on the gender and clothes of a person existing in an image captured by a camera by a simple method. The purpose is to provide.

  The present invention employs the following configuration in order to solve the above-described problems.

  That is, the information processing apparatus according to the present invention includes a storage unit for storing a human-shaped template and a clothing shape prepared for each type of clothing, and an edge detected in an image captured using a camera. Is compared with the template to determine whether the edge is the edge of a person included in the image, and for each person, the edge identified as the edge of the person included in the image is identified. A clothing shape prepared for each type of clothing, obtained as a human shape and estimating the sex of the person related to the human shape based on the shape characteristics that can be estimated from the human shape By comparing the person shape with the person shape, the type of clothes worn by the person related to the person shape is estimated, and based on the estimated sex and type of clothes, for each sex and type of clothes, Included in the image And a control unit which counts the number of people, characterized by comprising a.

  Here, the human shape refers to the shape of a human body, and may vary depending on the posture of the human and the imaged direction. The clothing shape refers to the shape of the clothing, and even the same clothing may vary depending on the posture of the person wearing the clothing and the imaged direction. Furthermore, the type of clothing is distinguished by the nature and form of the clothing such as the name of the clothing and the shape of the sleeve.

  According to the above configuration, the person shape of the person existing in the image captured using the camera is specified based on the matching between the edge detected in the image and the model of the person shape. Then, for each person, the gender of the person is estimated based on the shape features that can be estimated from the person shape. In addition, for each person, the type of clothes worn by the person is estimated based on the matching between the shape of the clothes prepared for each type of clothes and the shape of the person. Further, based on the estimated sex and type of clothing, the number of people included in the image is aggregated for each sex and type of clothing. In other words, according to the above configuration, the number of persons (gender and clothes information) of the gender and the type of clothing in the imaging range of the camera is tabulated based only on the shape included in the image captured using the camera. The

  Therefore, according to the above configuration, it is possible to acquire information on the gender and clothes of a person existing in an image captured by the camera by a simple method.

  As another form of the present invention, the control unit calculates a value related to a feature on the shape from which the gender can be estimated from the person shape, and compares the calculated value with a threshold value to calculate the person. The gender of the person related to the shape may be estimated.

  As another form of the present invention, the value related to the shape feature capable of estimating the gender may be a ratio using two lengths measured from the person shape.

  Here, the length measured from the human shape is, for example, a length, a head width, a shoulder width, an arm length, or the like in the human shape.

  According to the above configuration, the gender of the person related to the person shape is estimated by comparing the ratio using two lengths measured from the person shape with the threshold value. The ratio using the two lengths measured from the person shape indicates a relative relationship between the two lengths measured from the person shape, and thus does not depend on the size of the person shape.

  Therefore, according to the above configuration, the gender of the person related to the person shape can be estimated with the same threshold value regardless of the size of the person shape that varies depending on the distance between the camera and the person.

  As another form of the present invention, the human shape template may be a different template for each gender, including a shape feature that allows the gender to be estimated. Then, the control unit determines whether the edge detected in the image is identified as the edge related to the person shape by using a template related to which gender, so that the person related to the person shape is determined. Gender may be estimated.

  As another embodiment of the present invention, the control unit specifies an image capturing range based on information indicating an image capturing position of the image, and totals the number of people by sex and type of clothing in the image capturing range. May be.

As another form of the present invention, the control unit wears each type of clothing in the number of people for each gender based on the number of people counted for each sex and the type of clothing and the number of people for each sex. You may calculate the ratio of the number of people.

  According to the above configuration, the ratio of the number of people wearing each type of clothing in the number of people for each gender is calculated. In general, when a user understands some tendency, the ratio of the number of people to the total number of people is easier to understand than the number of people as it is.

  Therefore, according to the said structure, it becomes possible to provide the information which is easy to understand the tendency with respect to the kind of clothing worn for every sex.

  As another aspect of the present invention, an information processing method that implements each of the above configurations, a program, or a computer-readable storage medium that records such a program may be used. There may be. Further, as another aspect of the present invention, an information processing system in which a plurality of devices that realize each of the above configurations is configured to be communicable may be used.

  ADVANTAGE OF THE INVENTION According to this invention, the technique which makes it possible to acquire the information regarding the sex of the person who exists in the image imaged with the camera, and the clothing by a simple method can be provided.

The figure which illustrates the apparatus connection which concerns on embodiment. FIG. 3 illustrates an information processing device according to an embodiment. The figure which illustrates the record of the clothing template database which concerns on embodiment. 6 is a flowchart illustrating a processing example of the information processing apparatus according to the embodiment. The figure which illustrates the image acquired by the camera which concerns on embodiment. The figure which illustrates the edge detected in the image acquired by the camera which concerns on embodiment. The figure which illustrates the collation result with the edge detected in the image acquired with the camera which concerns on embodiment, and the model of a person shape. The figure which illustrates the gender estimation of the person which concerns on embodiment. The figure which illustrates the estimation of the clothing which concerns on embodiment. 6 is a diagram illustrating information provided by the information processing apparatus according to the embodiment. FIG.

  Hereinafter, embodiments of an information processing apparatus, an information processing method, a program, and the like according to one aspect of the present invention (hereinafter also referred to as “this embodiment”) will be described. However, the present embodiment is an exemplification, and the present invention is not limited to the configuration of the present embodiment.

  Although the data appearing in the present embodiment is described in a natural language (such as Japanese), more specifically, it is specified in a pseudo language, a command, a parameter, a machine language, or the like that can be recognized by a computer.

§1 Connection example of each device First, a configuration example of a network according to the present embodiment will be described. FIG. 1 shows a connection example of apparatuses according to the present embodiment. As shown in FIG. 1, the information processing apparatus 1 and the image acquisition apparatus 4 can communicate with each other via the Internet 5.

  The information processing apparatus 1 estimates the gender of a person present in an image captured using the camera 2 and the type of clothes worn by the person. Then, the information processing apparatus 1 adds up the number of people included in the image for each estimated gender and type of clothing.

  As shown in FIG. 1, the image acquisition device 4 includes a camera 2 and a communication unit 3. Further, the image acquisition device 4 has a hardware configuration and a functional configuration (not shown) known as a PC or a mobile phone.

  The camera 2 includes, for example, an image sensor that converts an image that is an optical signal into electricity, and records an image captured by the image sensor as digital data. The camera 2 is, for example, a street camera or a mobile terminal camera.

  A street camera is a camera installed on the street. When the camera 2 is a street camera, for example, the image acquisition device 4 is realized as a network camera, and the communication unit 3 is realized as a communication function of the network camera. In addition, for example, when the camera 2 is a street camera, the image acquisition device 4 is realized as a PC, and the communication unit 3 is realized as a communication function of the PC.

  When the camera 2 is a mobile terminal camera, the image acquisition device 4 is realized as a mobile phone, and the communication unit 3 is realized as a communication function of the mobile phone.

§2 Configuration example of information processing apparatus FIG. 2 illustrates the configuration of the information processing apparatus 1. As shown in FIG. 2, the information processing apparatus 1 has existing hardware such as a storage unit 11, a control unit 12, and an input / output unit 14 connected by a bus 13 as a hardware configuration.

  The storage unit 11 stores various data and programs used in processing executed by the control unit 12 (not shown). The storage unit 11 is realized by a hard disk, for example. The storage unit 11 may be realized by a recording medium such as a USB memory.

  The various data and programs stored in the storage unit 11 may be obtained from a recording medium such as a CD (Compact Disc) or a DVD (Digital Versatile Disc). The storage unit 11 may be called an external storage device.

The control unit 12 includes one or a plurality of processors such as a microprocessor or a CPU (Central Processing Unit), and peripheral circuits (ROM (Read Only Memory), RAM (Random Access Memory), interface circuits) used for processing of the processor. Etc.). The control unit 12 implements the processing of the information processing apparatus 1 in the present embodiment by executing various data and programs stored in the storage unit 11. ROM, RAM, and the like may be referred to as a main storage device in the sense that they are arranged in an address space handled by a processor in the control unit 12.

  The input / output unit 14 is one or a plurality of interfaces for transmitting / receiving data to / from a device existing outside the information processing device 1. The input / output unit 14 is known as a hardware interface such as an interface for connecting a LAN (Local Area Network) cable, an interface for connecting to a user interface, or a USB (Universal Serial Bus).

The input / output unit 14 transmits and receives IP (Internet Protocol) packets and the like to and from the image acquisition device 4 via the Internet 5. The input / output unit 14 may be connected to a user interface (not shown) (input / output devices such as a keyboard, a mouse, and a display). The input / output unit 14 may be connected to an input / output device such as a removable storage medium such as a CD drive or a DVD drive, or an input / output device such as a non-volatile portable storage medium such as a memory card. . Further, the input / output unit 14 may be called a communication unit in the sense of an interface for connecting to the Internet.

The information processing apparatus 1 may be configured by a general-purpose computer such as a PC (Personal Computer). Further, the information processing apparatus 1 may be called a server (for example, a WEB server) in the meaning of an apparatus that provides a service in response to a request from a user terminal (not shown).

  In the present embodiment, the information processing apparatus 1 estimates the gender of a person present in an image captured using the camera 2 and the type of clothes worn by the person. Then, the information processing apparatus 1 adds up the number of people included in the image for each estimated gender and type of clothing. The storage unit 11 and the control unit 12 that realize this process will be described below.

[Storage unit]
As shown in FIG. 2, the storage unit 11 includes a person template database 21 and a clothing template database 22. The person template database 21 and the clothing template database 22 are realized, for example, as data stored in a hard disk.

<Person template database>
The person template database 21 stores a person-shaped template (hereinafter also referred to as a “person-shaped template”). The person-shaped template is data for specifying an edge relating to a person existing in the image among edges detected in the image captured using the camera 2. The human shape template is expressed by, for example, graphic data indicating a human shape or a feature amount that can be calculated from the human shape. The feature amount is, for example, a SIFT (Scale-Invariant Feature Transform) or a texture feature amount.

  It should be noted that the human shape template represented in the present embodiment is a silhouette obtained by projecting a person in two dimensions. Here, the human shape may vary depending on the posture of the person and the state of the imaged direction. Therefore, the person-shaped template may be a two-dimensional silhouette reflecting the state of the person.

  Note that the person template database 21 may store one person shape template or a plurality of person shape templates. In another form of the present embodiment, the person-shaped template may be data that differs for each gender, including a shape feature that allows gender estimation.

<Clothing template database>
The clothing template database 22 stores clothing data. The clothing data is data including a clothing shape, and is prepared for each type of clothing. The clothing shape is data for estimating the type of clothing worn by a person present in an image captured using the camera 2. The clothing shape is expressed by, for example, graphic data indicating the shape of the clothing or a feature amount that can be calculated from the clothing shape, like the person-shaped template.

  In addition, the clothing shape in this embodiment expressed by these is a silhouette which projected clothing in two dimensions. Moreover, even if it is the same clothing, the clothing shape may vary depending on the posture of the person wearing the clothing and the imaged orientation. Therefore, the clothing shape in the present embodiment is a silhouette obtained by projecting the clothing in two dimensions reflecting the state of the clothing.

FIG. 3 shows a record example of the clothing template database 22 in the present embodiment. As shown in FIG. 3, the clothing data according to the present embodiment is realized as row data of records in the clothing template database 22. The clothing data according to the present embodiment includes a template ID field, a gender field, a wearing place field, a sleeve field, an orientation field, an item name field, and a graphic field.

  The template ID field stores information for distinguishing clothing data. As shown in FIG. 3, the information for distinguishing the clothing data is designated as “template 1150”, “template 1151”, or the like, for example.

  The sex field stores the sex information targeted by the clothes. The gender information targeted by the clothing is designated, for example, “male”, “female” or the like.

  The wear location field stores information indicating a location where the clothing is worn. For example, “upper body”, “lower body”, and the like are designated as the information indicating the location where the clothes are worn.

  The sleeve field stores information on the sleeve of the clothes. For example, “long sleeves”, “short sleeves”, “no sleeves”, and the like are designated as information on the sleeves of the clothes. In addition, about the clothing worn on the lower body, information on the length of the clothing is stored. The information on the length of the clothes is designated as “long”, “short”, or the like, for example.

  The direction field stores information indicating the direction of clothes included in the clothes data. The information indicating the direction of the clothing indicates the direction in which the clothing is imaged, and is designated, for example, “front”, “back”, “side”, or the like.

  Information indicating the item name of the clothing is stored in the item name field. The information indicating the article name of the clothing is designated as “short coat”, “jumper”, or the like, for example.

  The figure field stores clothing shapes or information for identifying clothing shapes. As described above, the clothing shape is, for example, image data or a feature amount, and the figure field may store the clothing shape as it is, or information for identifying the clothing shape is stored. Also good. The information for identifying the clothing shape is information that makes it possible to acquire the clothing shape based on the information, such as information indicating a storage area of the storage unit 11 in which the clothing shape is stored. .

  Note that the type of clothing in the present embodiment is indicated by using at least one of the information on the sleeve of the clothing stored in the sleeve field of the clothing data and the information indicating the product name of the clothing stored in the product name field of the clothing data. It is. The clothing data may include a field for storing information on the type of clothing in addition to the information on the sleeve of the clothing and the information indicating the name of the clothing shown in the present embodiment. In this case, the type of clothing is arbitrarily indicated from the information regarding the type of clothing.

[Control unit]
As shown in FIG. 2, the control unit 12 includes an image data acquisition unit 31, a person shape acquisition unit 32, a gender estimation unit 33, a clothing estimation unit 34, and a clothing tendency calculation unit 35. For example, the image data acquisition unit 31, the person shape acquisition unit 32, the gender estimation unit 33, the clothing estimation unit 34, and the clothing trend calculation unit 35 may include a program stored in the storage unit 11 in the peripheral circuit of the control unit 12. It is realized by being transferred to a certain RAM or the like and executed by the processor of the control unit 12.

In the present embodiment, the control unit 12 estimates the gender of a person present in an image captured using the camera 2 and the type of clothing worn by the person. And the control part 12 totals the number of persons included in an image according to the estimated sex and the kind of clothing. Hereinafter, processing of each component in the present embodiment will be briefly described. A description of specific processing of each configuration in the present embodiment will be described in “§3 Operation Example”.

<Image data acquisition unit>
The image data acquisition unit 31 acquires image data captured using the camera 2 from the image acquisition device 4. Further, the image data acquisition unit 31 specifies an image capturing range based on information indicating an image capturing position. Note that the image data may be either moving image data or still image data. The image data may be acquired by accessing the image acquisition device 4 from the information processing device 1 or by accessing the information processing device 1 from the image acquisition device 4.

<Human shape acquisition unit>
The person shape acquisition unit 32 detects an edge in the image acquired by the image data acquisition unit 31. Further, the person shape acquisition unit 32 collates the person shape template stored in the person template database 21 with the detected edge, and identifies whether the edge is the edge of the person included in the image. Then, the person shape acquisition unit 32 acquires an edge identified as the edge of the person included in the image as a person shape.

<Gender estimation unit>
The gender estimation unit 33 estimates the gender of the person related to the person shape based on the shape features that can be estimated from the person shape acquired by the person shape acquisition unit 32.

  In the present embodiment, the gender estimating unit 33 calculates a value related to the feature on the shape from which the gender can be estimated from the human shape acquired by the human shape acquiring unit 32. And the sex estimation part 33 estimates the sex of the person who concerns on this person shape by comparing this calculated value with a threshold value.

  Further, in the present embodiment, the value related to the shape feature for which gender can be estimated is a ratio using two lengths measured from the person shape acquired by the person shape acquisition unit 32. Here, the length measured from the human shape is, for example, a length corresponding to the height in the human shape acquired by the human shape acquisition unit 32, a length corresponding to the head width, a length corresponding to the shoulder width, The length corresponds to the length of the arm.

  Further, in the present embodiment, the human shape template stored in the human template database 21 may be different data for each gender, including features on the shape for which gender can be estimated. At this time, the gender estimation unit 33 determines in the human shape acquisition unit 32 whether the edge detected in the image is identified as the edge related to the human shape by using the human shape template related to any gender. Thus, the gender of the person related to the person shape may be estimated.

<Clothing estimation part>
The clothing estimation unit 34 collates the clothing shape included in the clothing data stored in the clothing template database 22 with the person shape acquired by the person shape acquisition unit 32, so that the person related to the person shape wears it. Estimate the type of clothing you have.

<Clothing trend calculator>
The clothing tendency calculation unit 35 is included in the image acquired by the image data acquisition unit 31 for each sex and type of clothing based on the sex and the type of clothing estimated by the sex estimation unit 33 and the clothing estimation unit 34. Count the number of people. In addition, the clothing tendency calculation unit 35 calculates the ratio of the number of people wearing each type of clothing in the number of people for each gender based on the number of people for each gender and the type of clothing and the number of people for each gender. .

§3 Operation Example Next, an operation example in the present embodiment will be described with reference to FIGS. 4, 5, 6A, 6B, 7, 8, and 9. FIG. The flowchart of FIG. 4 illustrates the process of the information processing apparatus 1 according to the present embodiment. FIG. 5 illustrates an image acquired by the camera 2 according to the present embodiment. FIG. 6A illustrates a state where an edge in the image shown in FIG. 5 is detected in the human shape acquisition unit 32 according to the present embodiment. FIG. 6B exemplifies a matching result between the edge shown in FIG. 6A and the person-shaped template in this embodiment. FIG. 7 illustrates the operation of the gender estimation unit 33 in the present embodiment. FIG. 8 illustrates the operation of the clothing estimation unit 34 in the present embodiment. FIG. 9 illustrates information provided by the information processing apparatus 1 in the present embodiment.

<Start>
First, as illustrated in FIG. 4, the information processing apparatus 1 starts processing by executing, for example, a program stored in the storage unit 11 by the control unit 12 in a predetermined cycle (S100).

<Acquisition of image data>
When the process is started, the image data acquisition unit 31 acquires image data from the image acquisition device 4 using the camera 2 (S101). FIG. 5 illustrates the acquired image data. In this operation example, the image data acquisition unit 31 specifies the imaging range and imaging date and time of the acquired image. The camera 2 is, for example, a street camera or a mobile terminal camera.

  When the camera 2 is a street camera, the image acquisition device 4 is, for example, a network camera or a PC. When the image acquisition device 4 is a network camera or a PC, the image acquisition device 4 accesses the Internet 5 by the communication unit 3 based on a request from the image data acquisition unit 31 or a spontaneous operation, etc. The image data picked up and transmitted is transmitted to the information processing apparatus 1.

  The image data acquisition unit 31 accesses the image acquisition device 4 and waits until image data is received from the image acquisition device 4, or receives image data from the image acquisition device 4 and is already stored in the storage unit 11. To obtain image data captured by the street camera.

  Note that the image data transmitted from the image acquisition apparatus 4 to the information processing apparatus 1 and captured using a street camera includes the identification number of the street camera as its attribute information.

  The image data acquisition unit 31 specifies the position of the street camera using the identification number of the street camera. For example, the storage unit 11 stores information indicating the position of the street camera associated with the street camera identification number. And the image data acquisition part 31 collates with the information which shows the position of the street camera stored in the memory | storage part 11 using the identification number of the street camera contained in the received image data, and transmitted the received image data Locate the street camera.

  The image data acquisition unit 31 also indicates the date and time when the image acquisition device 4 was accessed, the date and time when the image data was received from the image acquisition device 4, or the date and time when the image data received from the image acquisition device 4 was stored in the storage unit 11. The image data is identified as the imaging date and time.

On the other hand, when the camera 2 is a mobile terminal camera, the image acquisition device 4 is, for example, a mobile phone. When the image acquisition device 4 is a mobile phone, the image acquisition device 4 normally accesses the Internet 5 by the communication unit 3 based on an operation by the user of the image acquisition device 4 and takes an image using the camera 2. The image data is transmitted to the information processing apparatus 1. That is, the transmission of the image data is generally performed at an arbitrary timing desired by the user of the mobile terminal camera.

  The image data acquisition unit 31 waits until reception of image data transmitted at the arbitrary timing, or is captured by the mobile terminal camera from image data that has already been received and stored in the storage unit 11. Acquired image data.

It should be noted that image data transmitted from the image acquisition device 4 to the information processing device 1 and imaged using the mobile terminal camera includes, as its attribute information, the imaging date and time and GPS (Global Positioning System) information by the mobile terminal camera. It is included. Further, the image data transmitted from the mobile terminal camera to the information processing apparatus 1 includes the user name of the mobile terminal camera and a comment input by operating the mobile terminal camera as the attribute information. It may be. The user of the mobile terminal camera is a person who provides image data to the information processing apparatus 1 and is hereinafter referred to as a contributor.

  The image data acquisition unit 31 specifies the position of the mobile terminal camera using the GPS information included in the image data. Further, the image data acquisition unit 31 specifies the imaging date / time included in the image data as the imaging date / time of the image data.

  Note that the information indicating the position of the street camera and the GPS information correspond to information indicating the imaging position of the image of the present invention.

  The image data acquisition unit 31 specifies the position of the camera 2 by these processes. Then, the information processing apparatus 1 specifies the imaging range and imaging date / time of the image captured by the camera 2 based on the identified position of the camera 2. Note that the information obtained by specifying the imaging range may be rough information such as “XX street XX near the 1-chome intersection”, for example.

<Edge detection>
Next, as shown in FIG. 4, when the image data is acquired, the human shape acquisition unit 32 detects edges in the image acquired by the image data acquisition unit 31 (S102). FIG. 6A illustrates the edges detected by the processing.

  Edge detection is performed by a well-known method. The person shape acquisition unit 32 detects an edge in an image by a process using, for example, a Sobel filter or a Prewitt filter.

<Acquisition of person shape>
Next, as shown in FIG. 4, the human shape acquisition unit 32 collates the human shape template stored in the human template database 21 with the detected edge, and the edge of the person included in the image is detected. Identify whether it is an edge. Then, the human shape acquisition unit 32 acquires an edge identified as the human edge included in the image as a human shape (S103). FIG. 6B illustrates the human shape detected by the processing.

  As described above, the human shape template stored in the human template database 21 is, for example, graphic data indicating a human shape or a feature amount that can be calculated from the human shape.

  For example, when the human shape template is graphic data, the human shape acquisition unit 32 performs pattern matching between the graphic data and the detected edge, and specifies an edge that can match or approximate the graphic data.

  When the human shape template is a feature amount, the human shape acquisition unit 32 calculates the feature amount from the detected edge. The calculated feature value is also called a local feature value in the sense that it is a feature value obtained by cutting out a part of an image. Then, the human shape acquisition unit 32 compares the feature amount calculated from the edge with the human shape template (feature amount), and specifies an edge that can match or approximate the human shape template.

  Here, it is considered that the edges of the person included in the image include an edge corresponding to the outline of the clothes and an edge existing in the clothes. The edge inherent in the clothes is an edge of a tool, a design, or the like provided in the clothes such as a button, a fastener, or a pattern.

  On the other hand, a person-shaped template is data that takes into account the edges inherent in clothes, that is, data that is templated to buttons, fasteners, patterns, and other equipment or designs on clothes, in addition to the outline of a person. Not always.

  For example, when the human-shaped template is not data that takes into account the edges inherent in the clothes, the edges specified as being able to match or approximate the human-shaped template by the above-described processes are usually edges existing in the clothes. Is not included. However, even in this case, the edges identified as being able to match or approximate the person-shaped template by the above processes include the edges corresponding to the outline of the clothes.

  Therefore, first, the person shape acquisition unit 32 specifies the edge specified as being able to match or approximate the person shape template by each of the above processes as the edge corresponding to the outline of the clothing. In addition, the person shape obtaining unit 32 identifies an edge existing in the outline of the clothing specified by the edge corresponding to the outline of the clothing as an edge existing in the clothing. Then, the person shape acquisition unit 32 identifies the edge specified by each of the above processes, that is, the edge corresponding to the outline of the clothes and the edge existing in the outline of the clothes from the edges of the person included in the image. To do. Then, the person shape acquisition unit 32 acquires a set of these edges as a person shape.

  On the other hand, when the person-shaped template is data that takes into account the edges inherent in the clothes, the edges identified as being able to match or approximate the person-shaped template by the above-described processing include edges corresponding to the outline of the clothes And edges inherent in the clothes. Therefore, the person shape acquisition unit 32 identifies these edges as the edges of the person included in the image. Then, the person shape acquisition unit 32 acquires a set of these edges as a person shape.

<Estimating the gender of a person>
Next, as shown in FIG. 4, the gender estimation unit 33 relates to the person shape based on the shape features that can be estimated from the person shape acquired by the person shape acquisition unit 32. The gender of the person is estimated (S104).

In the present embodiment, the gender estimation unit 33 first calculates a value related to the feature on the shape from which the gender can be estimated from the human shape acquired by the human shape acquisition unit 32. The value related to the shape feature that can estimate the gender is a ratio using two lengths measured from the person shape acquired by the person shape acquisition unit 32. Here, an example of the length measured from the person shape is shown in FIG. The length measured from the human shape corresponds to, for example, a length corresponding to the height related to the human shape (arrow 50 in FIG. 7), a length corresponding to the head width (arrow 51 in FIG. 7), and a shoulder width. It is the length (arrow 52 in FIG. 7) or the length corresponding to the length of the arm (arrow 53 in FIG. 7). The human shape acquisition unit 32 measures at least two lengths measured from these human shapes, and uses, for example, at least two measured lengths, for example, a shoulder width with respect to height, or
Calculate the ratio of head width to shoulder width.

  And the sex estimation part 33 estimates the sex of the person who concerns on this person shape by comparing this calculated value with a threshold value. The threshold is set in advance by a program or the like stored in the storage unit 11. For example, the median of the average value of male ratio and the average value of female ratio, etc., which can be calculated using the average height of each gender published by the National Health Survey conducted by the Ministry of Health, Labor and Welfare, etc. Are set as the threshold values.

  In the present embodiment, the person-shaped template stored in the person template database 21 may be different data for each gender, including features on the shape for which gender can be estimated. At this time, the gender estimation unit 33 determines in the human shape acquisition unit 32 whether the edge detected in the image is identified as the edge related to the human shape by using the human shape template related to any gender. Thus, the gender of the person related to the person shape is estimated.

  For example, the ratio of the average shoulder width to the average height and the ratio of the average head width to the average shoulder width of each male and female is applied to the data that differs by gender, including the shape features that can estimate gender. It is a template with a different person shape for each gender. In other words, the person-shaped template in another form of the present embodiment is data including physical characteristics regarding a shape that is different for each gender and is prepared for each gender.

<Estimating the clothes of a person>
Next, as shown in FIG. 4, the clothing estimation unit 34 collates the clothing shape included in the clothing data stored in the clothing template database 22 with the person shape acquired by the person shape acquisition unit 32, The type of clothes worn by the person related to the person shape is estimated (S105). FIG. 8 illustrates the process.

  As shown in FIG. 8, in the present embodiment, the clothing estimation unit 34 estimates the type of clothing worn by a person for each place where the clothing is worn (upper body and lower body). Accordingly, in the present embodiment, the clothing estimation unit 34 divides the clothing shape of the clothing worn on the upper body and the clothing shape of the clothing worn on the lower body, and collates them with the person shape.

  Here, the matching of each clothing shape and the person shape can be explained in the same manner as the matching of the person shape template and the edge existing in the image, which is performed by the person shape acquisition unit 32. Therefore, the description is omitted. On the other hand, the verification performed by the person shape acquisition unit 32 is different in that different data is used for the upper body and the lower body. This difference will be described below.

  First, the clothing estimation unit 34 acquires from the storage unit 11 a clothing shape for collation with each of the human shapes acquired by the human shape acquisition unit 32. At this time, the gender of the person related to the person shape is estimated by the process of the gender estimation unit 33 (S104). Therefore, the clothing estimation unit 34 acquires the clothing shape from the storage unit 11 based on the gender and the wearing location estimated for each person shape.

  For example, it is assumed that the gender of the person related to the person shape is male, and the clothing estimation unit 34 estimates the clothing of the upper body of the person. At this time, the clothing estimation unit 34 acquires clothing data in which “male” is stored in the gender field and “upper body” is stored in the wearing location field from the clothing template database 21. Then, the clothing estimation unit 34 acquires the clothing shape stored or shown in the graphic field included in the acquired clothing data from the storage unit 11.

  Next, the clothing estimation unit 34 compares the clothing shape acquired for each person's gender related to the person's shape and the clothing wear location with the person shape, and identifies a clothing shape that matches or approximates the person shape for each wear location. Then, the clothing estimation unit 34 uses the information on the type of clothing stored in the clothing data related to the clothing shape identified by the collation for the type of clothing worn by the person related to the person shape for each wearing location. presume.

  In addition, the method of collating a clothing shape with a person shape may be arbitrary. In the present embodiment, as shown in FIG. 8, the clothing estimation unit 34 determines, from a person shape, an edge related to clothing worn by a person on the upper body and an edge related to clothing worn by a person on the lower body. Identify. And the clothing estimation part 34 collates this specified edge (person shape) and clothing shape for every wear location. Thus, the clothing estimation unit 34 specifies a clothing shape that can match or approximate the edge related to the clothing worn on the upper body and a clothing shape that can match or approximate the edge related to the clothing worn on the lower body. Then, the clothing estimation unit 34 estimates the type of clothing worn by the person related to the person shape based on the information about the type of clothing stored in the clothing data related to the clothing shape specified in the upper body and the lower body.

  The outline of the clothing is divided into an upper body and a lower body, except in special cases such as when the clothing being worn is a long coat. For example, based on the outline of the clothing that separates the upper body and the lower body, the clothing estimation unit 34 applies an edge relating to the clothing worn by the person to the upper body and the clothing worn by the person on the lower body from the shape of the person. Such an edge is specified.

  In addition, as an example of a method for matching the clothing shape against the human shape, a method using a ratio related to the human shape region is conceivable. That is, the clothing estimation unit 34 assumes that the upper part of the person shape is an upper body with respect to the length from the upper end to the lower end of the person shape with the head side of the person as the upper part, and the lower part of the person shape. Assuming a predetermined proportion of the region as the lower body, the clothing shape may be collated with each region.

  At this time, for example, the clothing estimation unit 34 assumes that the upper 55% region of the human shape acquired by the human shape acquisition unit 32 is the upper body region and the lower 45% region is the lower body region. Then, the clothing estimation unit 34 collates the clothing shape relating to the upper body with the human shape relating to the region assumed to be the upper body. Also, the clothing estimation unit 34 collates the clothing shape relating to the lower body with the human shape relating to the region assumed to be the lower body.

  Note that the region assumed to be the upper body and the region assumed to be the lower body may not be completely separated. That is, the region assumed to be the upper body and the region assumed to be the lower body may overlap.

  In addition, as an example of a method for collating the clothing shape with the human shape, a method of collating the human shape with the clothing shape using a predetermined position in the human shape as a reference for the wearing part is conceivable. That is, the clothing estimation unit 34 may collate each clothing shape with the person shape with the head side of the person as the upper part and a predetermined position of the person shape as a reference for the wearing part.

  At this time, for example, the clothing estimation unit 34 sets the length from the upper end to the lower end of the human shape acquired by the human shape acquisition unit 32 as 100%, and the position below the human shape ratio 12% from the upper end of the human shape. May be used as a reference for the upper body. Then, the clothing estimation unit 34 may collate the clothing shape with the human shape by matching the upper end of the clothing shape of the upper body with the reference position. The set value of 12% is a value set in consideration that the region related to the ratio in the human shape is a region corresponding to the head of the human shape, and can be arbitrarily set.

  For example, the clothing estimation unit 34 sets the length from the upper end to the lower end of the human shape acquired by the human shape acquisition unit 32 as 100%, and the person from the lower end of the human shape acquired by the human shape acquisition unit 32. A position that is 5% above the shape ratio may be used as a reference for the lower body. Then, the clothing estimation unit 34 may collate the clothing shape with the human shape by aligning the lower end of the clothing shape relating to the lower body with the reference position. The set value of 5% is a value set in consideration that the area related to the ratio in the human shape is an area corresponding to the foot of the lower limbs of the human shape, and can be arbitrarily set. .

  Further, as an example of a method for collating the clothing shape with the human shape, a method of repeatedly performing the collation while shifting the position where the clothing shape is collated with respect to the human shape is conceivable. That is, the clothing estimation unit 34 may collate the clothing shape with the human shape while shifting the upper end of the clothing shape downward at predetermined intervals from the upper end of the human shape with the head side of the person as the upper part.

  The clothing estimation unit 34 may collate the clothing shape with the human shape while shifting the lower end of the clothing shape upward at predetermined intervals from the lower end of the human shape with the head side of the person as the upper part.

  Further, as an example of a method for matching the clothing shape against the human shape, a method using region determination information for the human shape is conceivable. That is, the upper body region and the lower body region may be determined in the person shape. Then, the clothing estimation unit 34 may collate the clothing shapes with the determined areas. The determination of the area related to the person shape is performed, for example, during the process (S103) of the person shape acquisition unit 32.

  At this time, for example, an upper body region and a lower body region are set in a person-shaped template stored in the person template database 21. Then, after identifying an edge that can match or approximate the human shape template, the human shape acquisition unit 32 determines that the edge related to the upper body region set in the human shape template is the upper body region in the human shape. To do. Further, the person shape acquisition unit 32 determines an edge relating to the lower body region set in the human shape template as the lower body region in the person shape.

  It should be noted that the person-shaped template in this embodiment may be a two-dimensional silhouette that reflects the state of the person, the captured direction, and the like. At this time, the clothing estimation unit 33 determines whether the edge detected in the image is identified as the edge related to the human shape by the human shape template reflecting any human state in the human shape acquisition unit 32. You may judge. And the clothing estimation part 33 may acquire the information regarding the state of the person who concerns on this person shape by the said determination. The information related to the state of the person can include information such as the posture of the person related to the person shape and the imaged direction. The clothing estimation unit 33 may use the information when obtaining clothing data from the clothing template database 22. For example, when the information on the person's state includes information on the direction in which the person related to the person shape is captured, the clothing estimation unit 33 uses the information on the direction in which the person related to the person shape is captured in the direction field. Clothing data that matches the stored information may be obtained from the clothing template database 22.

<Total number of people>
Next, as shown in FIG. 4, the clothing trend calculation unit 35 acquires image data for each gender and type of clothing based on the gender and the type of clothing estimated by the gender estimation unit 33 and the clothing estimation unit 34. The number of persons included in the image acquired by the unit 31 is totaled (S106).

The clothing tendency calculation unit 35 counts each time the clothing estimation unit 34 estimates the type of clothing in the shape of a person, so that it is included in the image acquired by the image data acquisition unit 31 for each gender and type of clothing. You may count the number of people. That is, the process of S106 may be executed in parallel with the process of S105.

  Here, as described above, the type of clothing in this embodiment is at least one of the information on the sleeve of the clothing stored in the sleeve field of the clothing data and the information indicating the product name of the clothing stored in the product name field of the clothing data. It is specified by using.

  The clothing tendency calculation unit 35 may prepare a plurality of “clothing types” and count the number of each “clothing type” separately for each gender and clothing type in the imaging range of the camera. For example, in the present embodiment, the clothing trend calculation unit 35 may separately perform aggregation based on information about clothes sleeves and aggregation based on information indicating the names of clothes (see FIG. 9 described later).

<Calculation ratio>
Next, the clothing trend calculation unit 35 calculates the ratio of the number of people wearing each type of clothing in the number of people for each gender based on the number of people counted for each gender and type of clothing and the number of people for each gender. (S107).

  First, the clothing tendency calculation unit 35 obtains the total number (the total number of men and the total number of women) for each gender of the person in the image and related to the person shape acquired by the person shape acquisition unit 32.

  For example, the clothing tendency calculation unit 35 obtains the total number of persons for each gender based on gender estimation by the gender estimation unit 33. This process may be performed by the gender estimating unit 33. That is, in the process of S104, each time the gender estimating unit 33 estimates the gender of the person related to the person shape acquired by the person shape acquiring unit 32, the number of persons estimated to be male and the number of persons estimated to be female are By counting, the total number of persons for each gender may be obtained.

  And the clothing tendency calculation part 35 is the number of people who wear each kind of clothes, for example by dividing the number of persons totaled according to the sex and the kind of clothes calculated | required in the process of S106 by the total number of the persons for every sex. Calculate the ratio of the number of people for each gender.

  Note that the calculation of the ratio may be performed for a plurality of clothing types, as in the above “<Total number of people>”.

<End>
Finally, for example, the control unit 12 obtains information on the ratio of the number of people wearing each type of clothing calculated for each gender and the number of people counted by sex and the type of clothing obtained by the clothing trend calculation unit 35. Are stored in the storage unit 11 in association with the image data. Thereby, the information processing apparatus 1 ends the process (S108).

  Note that the control unit 12 also associates information on the imaging range and imaging date and time specified by the image data acquisition unit 31 with the image data. Accordingly, it is possible to perform a search for specifying at least one of the imaging range and the imaging date / time for the image data stored in the storage unit 11. And the user wears each type of clothing calculated for each gender and the number of people counted in the gender and the type of clothing in the image specified by at least one of the imaging range and the imaging date and time. It becomes possible to know the information of the ratio of the number of people who are doing. Hereinafter, this information is referred to as clothing trend information.

When the image data acquired by the image data acquisition unit 31 is image data captured by a mobile terminal camera, the attribute information of the image data may include a user name and a comment. The control unit 12 may store the user name and comment information as they are in the storage unit 11 as image data attribute information. Thereby, the user can know the comment of the poster in addition to the information on the clothing tendency.

§4 Actions and effects according to the embodiment As described above, in the information processing apparatus 1 according to the present embodiment, first, the image data acquired by the image data acquisition unit 31 using the camera 2 includes the imaging range and the imaging date and time. Is identified and acquired. Then, the human shape acquisition unit 32 acquires a human shape from the acquired image. Subsequently, the gender of the person related to the person shape acquired by the person shape acquisition unit 32 is estimated by the gender estimation unit 33 by using the shape features that can be estimated from the person shape. In addition, the type of clothing worn by the person related to the person shape acquired by the person shape acquisition unit 32 is estimated by collating the clothing shape and the person shape by the clothing estimation unit 34. Furthermore, the clothing tendency calculation unit 35 calculates the clothing tendency (information on the gender and clothes of the person).

  Therefore, in this embodiment, since the clothing tendency is calculated only by collation and comparison based on the shape, information on the gender and clothing of the person existing in the image captured by the camera 2 is acquired by a simple method. It is possible.

§5 Modifications Embodiments of the present invention have been described in detail above. However, the above description is merely an example of the present invention in all respects and is not intended to limit the scope thereof. It goes without saying that various improvements and modifications can be made without departing from the scope of the present invention.

  For example, the information on the clothing tendency associated with the image data and stored in the storage unit 11 may be provided to the user terminal via the Internet 5.

  At this time, the user of the user terminal designates the location and the date and time via the user terminal interface. Based on the designation, the user terminal requests the information processing apparatus 1 to provide information on the clothing trend specifying the location and the date and time.

  When the information processing apparatus 1 receives this request, the information processing apparatus 1 acquires clothes trend information from the storage unit 11 based on the designated location and date. Then, the information processing apparatus 1 returns the acquired clothes tendency information to the user terminal. In addition, when there is information on the user name and comment of the poster, the information processing apparatus 1 returns these information in accordance with the clothing trend information. FIG. 9 is an example of information displayed on a user terminal that has received a clothing trend. In the example shown in FIG. 9, the imaging range, imaging date and time, the ratio of the number of people wearing each type of clothing to the number of people for each gender (based on information about clothing sleeves, and information indicating the name of clothing) The user name and comment of the poster.

  Further, for example, since the imaging range and the imaging date / time are specified, these pieces of information may be provided together with information providing services such as map information, public institution transfer information, weather forecast information, or travel information. .

  Thereby, in this modification, it is possible to provide information on clothing trends at a specific point to a user who intends to go to the specific point. Moreover, in this modification, the user can make his / her life comfortable through the information on the clothing tendency.

§6 Supplement The scope of the present invention is interpreted only by the claims. Moreover, those skilled in the art can implement an equivalent range from the description of the present embodiment based on the description of the claims and the common general technical knowledge. The terms used in this specification are as follows unless otherwise specified.
Used in the meaning normally used in the field. Thus, unless defined otherwise, all technical and technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. In the event of a conflict, terms used herein will be understood in the meaning set forth herein (including definitions).

DESCRIPTION OF SYMBOLS 1 Information processing apparatus 2 Camera 3 Communication part 4 Image acquisition apparatus 5 Internet 11 Memory | storage part 12 Control part 13 Bus 14 Input / output part 21 Person template database 22 Clothing template database 31 Image data acquisition part 32 Person shape acquisition part 33 Gender estimation part 34 Clothing estimation unit 35 Clothing tendency calculation unit

Claims (8)

A storage unit for storing a person-shaped template and a clothing shape prepared for each type of clothing;
Identifying whether the edge is the edge of a person included in the image by matching an edge detected in the image captured using the camera with the template,
For each person, an edge identified as the edge of the person included in the image is acquired as a person shape,
Estimating the gender of the person related to the person shape based on the shape characteristics that can be estimated from the person shape, gender,
By comparing the clothing shape prepared for each type of clothing with the person shape, estimating the type of clothing worn by the person related to the person shape,
Based on the estimated gender and type of clothing, a control unit that counts the number of people included in the image for each gender and type of clothing,
An information processing apparatus comprising:   The control unit calculates a value related to a shape feature capable of estimating the sex from the person shape, and estimates the gender of the person related to the person shape by comparing the calculated value with a threshold value. Item 4. The information processing apparatus according to Item 1.   The information processing apparatus according to claim 2, wherein the value related to the feature on the shape capable of estimating the gender is a ratio using two lengths measured from the person shape. The human-shaped template is a different template for each gender, including a shape feature that allows the gender to be estimated.
The control unit determines the gender of the person related to the person shape by determining whether the edge detected in the image is identified as the edge related to the person shape by using a template related to which gender. The information processing apparatus according to claim 1, wherein estimation is performed.   The control unit identifies an image capturing range based on information indicating an image capturing position of the image, and totals the number of persons according to the sex and the type of clothing in the image capturing range. The information processing apparatus according to any one of the above.   The control unit calculates a ratio of the number of people wearing each type of clothing in the number of people for each gender based on the number of people for each sex and the number of people for each sex. The information processing apparatus according to any one of claims 1 to 5. A computer having a storage unit for storing a figure of a person shape and a clothing shape prepared for each type of clothing,
Identifying whether the edge is the edge of a person included in the image by matching an edge detected in an image captured using a camera with the template;
Obtaining for each person the edge identified as the edge of the person included in the image as a person shape;
Estimating a gender of a person related to the person shape based on a shape feature that can be estimated from the person shape and gender;
Estimating the type of clothing worn by a person related to the person shape by comparing the shape of the clothing prepared for each type of clothing with the person shape;
Based on the estimated gender and type of clothing, totaling the number of people included in the image for each gender and type of clothing;
The information processing method characterized by performing. A computer equipped with a storage unit for storing a person-shaped template and a clothing shape prepared for each type of clothing,
Identifying whether the edge is the edge of a person included in the image by matching an edge detected in an image captured using a camera with the template;
Obtaining for each person the edge identified as the edge of the person included in the image as a person shape;
Estimating a gender of a person related to the person shape based on a shape feature that can be estimated from the person shape and gender;
Estimating the type of clothing worn by a person related to the person shape by comparing the shape of the clothing prepared for each type of clothing with the person shape;
Based on the estimated gender and type of clothing, totaling the number of people included in the image for each gender and type of clothing;
A program for running

Images