JP5293636B2 - Model image creation device and model image creation program for model image creation device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a model image creation device for creating a model image in which only clothing and ornament of a person is modeled, and a model image creation program of the model image creation device. <P>SOLUTION: When parts of the person are extracted from the whole body image data (S15). Person image data is extracted based on the extracted parts of the person (S20). Clothing and ornament data is segmented from the person image data, and matching of the segmented clothing and ornament data with a clothing and ornament template is performed. The shape of the approximate template is segmented from the person image data, and clothing and ornament synthetic data synthesized with the template is created. Clothing and ornament avatar data is created based on a brightness pattern of the clothing and ornament synthetic data. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to a model image creation device and a model image creation program for a model image creation device.

  2. Description of the Related Art Conventionally, services that provide model images that reflect themselves, called avatars, are used in various scenes such as online games and karaoke. The user can freely operate by himself such as moving an avatar or changing clothes and decorations (hereinafter referred to as clothing). By using such an avatar, communication on the network is enriched. However, the avatar's apparel was selected from parts provided in advance by a company, and only a combination of them was used, and the user could not edit it. Therefore, there has been a demand for a technique capable of extracting only clothes from a person, modeling it, and editing the data.

  For example, in Patent Document 1, a three-dimensional human body model that reflects the morphological features of each individual is extracted by extracting the size of a human body part that is a trait individual feature amount using only image processing from a plurality of captured images. An automatic human body copy model creation device has been proposed. This automatic human body copy model creation apparatus obtains a deformation parameter by referring to the standard size of the human body based on the extracted size of the human body part. Then, by transforming any given basic model, a three-dimensional human body model that reflects the morphological characteristics of each individual is generated while maintaining a unified character as a whole.

JP 2001-312743 A

  However, the automatic human body copy model creation apparatus described in Patent Document 1 has a problem that not only clothing but also extraneous parts such as the head and limbs are modeled, so that only clothing can not be modeled.

  The present invention has been made to solve the above-described problems, and provides a model image creation device that can create a model image that models only the clothing of a person, and a model image creation program for the model image creation device. Objective.

  The model image creation device according to the first embodiment of the present invention is a model image creation device that creates clothing model image data, which is an image obtained by modeling the shape of a person's clothing, for each type of shape of the person's clothing. A block that associates a partial block corresponding to each part of the human body with the clothing image pattern storage unit that stores the clothing image pattern that is the image pattern of the image and the image data of the person imaged by the imaging unit that images the person A clothing data extraction unit that extracts clothing data that is data of a clothing part from the partial block associated by the block association unit; and the clothing data extracted by the clothing data extraction unit; By comparing the clothing image pattern stored in the clothing image pattern storage means, the clothing Determine the clothing image pattern that approximates data, and based on the clothing image pattern, the clothing shape specifying means for specifying the shape of the person's clothing, and the shape of the clothing specified by the clothing shape specifying means, Clothing model image data creation means for creating clothing model image data that is model image data of a person's clothing, and clothing model image data output means for outputting the clothing model image data created by the clothing model image data creation means And.

  In the first embodiment of the present invention, the person is imaged by the imaging means. The clothing image pattern storage means stores a clothing image pattern that is an image pattern for each type of clothing shape. The block association means associates partial blocks corresponding to each part of the human body with the image data of the person imaged by the imaging means. The clothing data extraction unit extracts the clothing data that is the data of the clothing part from the partial blocks associated by the block association unit. The clothing shape specifying means determines the clothing image pattern that approximates the clothing data by comparing the clothing data extracted by the clothing data extracting means and the clothing image pattern stored in the clothing image pattern storage means, The shape of the person's clothing is specified based on the image pattern. The clothing model image data creating means creates clothing model image data that is data of a model image of a person's clothing based on the shape of the clothing specified by the clothing shape specifying means. The clothing model image data output means outputs the clothing model image data created by the clothing model image data creation means. Thereby, one's own clothes can be reflected as clothes of a model image as they are. Furthermore, since the pattern patterns of model images are rich, a variety of model images can be created.

  In the first embodiment, the image capturing data of the person includes design extraction means for extracting a design including a clothing pattern from a part corresponding to the shape of the clothing specified by the clothing shape specifying means, The clothing model image data creation means may create the clothing model image data by synthesizing the shape of the clothing identified by the clothing shape identification means and the design extracted by the design extraction means. . The design extraction unit extracts a design including a clothing pattern from a part corresponding to the shape of the clothing specified by the clothing shape specifying unit in the human image data. The clothing model image data creation unit creates the clothing model image data by synthesizing the shape of the clothing identified by the clothing shape identification unit and the design extracted by the design extraction unit. Therefore, the actual design of the clothing can be reflected in the shape of the clothing.

  In the first embodiment, the clothing data extraction unit includes a skin color region specifying unit that specifies a skin color region in the partial block associated by the block association unit, and is specified by the skin color region specifying unit. You may make it extract the area | region remove | excluding the said skin color area | region from the said partial block as said clothing data. The skin color area specifying unit specifies the skin color area in the partial block associated by the block association unit. The skin color area is a portion of the person image other than the clothing such as the head and limbs. By removing such a region from the partial block associated with the image data of the person imaged by the imaging means, only the clothing data can be extracted.

  In the first embodiment, the skin color area specifying unit extracts the skin color area from a partial block corresponding to the face part of the person among the partial blocks associated with the imaging data by the block association unit. A reference value for specifying may be acquired, and the skin color region may be specified based on the reference value. The face portion is the most exposed portion in the human image data. By acquiring the reference value in the partial block corresponding to the face portion and specifying the skin color area with the reference value, the clothing portion can be reliably extracted.

  Moreover, in the first embodiment, the clothing shape specifying means follows the shape of the clothing data pattern stored in the clothing image pattern storage means along the shape of the clothing data extracted by the clothing data extraction means. You may make it provide the shape change means to change, and make it compare the said clothing data extracted by the said clothing data extraction means, and the said clothing image pattern from which the shape was changed by the said shape change means. The shape of the clothing data is not constant depending on the pose of the person in the image data of the person. Therefore, the shape changing unit changes the shape of the clothing image pattern stored in the clothing image pattern storage unit along the shape of the clothing data extracted by the clothing data extracting unit. Thereby, since the clothing shape specification means can compare the clothing data and the clothing image pattern, the clothing shape can be easily specified.

  In the first embodiment, there may be provided clothing model image data changing means for changing the shape and design of the clothing model image corresponding to the clothing model image data created by the clothing model image data creating means. The clothing model image data changing unit can change the shape and design of the clothing model image corresponding to the clothing model image data created by the clothing model image data creating unit. Therefore, the clothing model image can be freely changed according to the user's preference. It is also possible to correct the unclear part of the clothing model image.

  Further, in the first embodiment, block position changing means for changing the position of the partial block associated by the block associating means may be provided. When the position of the partial block associated by the block association unit is not appropriate, the position of the partial block associated by the block association unit can be changed by the block position changing unit. Thereby, the clothing data can be reliably extracted from the partial block.

The model image creation program of the model image creation device according to the second embodiment of the present invention is a model image creation device that causes a model image creation device to create clothing model image data that is an image obtained by modeling the shape of a person's clothing. And a block association step of associating a partial block corresponding to each part of the human body with the image data of the person imaged by the imaging means for imaging a person, and the block association Apparel data extraction step for extracting apparel data, which is data of an apparel part, from the partial block associated in the step, the apparel data extracted in the apparel data extraction step, and an image pattern for each type of apparel shape A clothing image pattern storage hand that stores a clothing image pattern A clothing shape specifying step that determines the clothing image pattern that approximates the clothing data by comparing the clothing image pattern stored in the clothing data, and identifies the shape of the person's clothing based on the clothing image pattern; based on the clothing of the shape specified in the clothing shape specifying step, a clothing model image data creation step of creating a fashion model image data is data of a model image by modeling the clothing of the person, the clothing model image data A model image creation device executes the clothing model image data output step of outputting the clothing model image data created in the creation step.

In the second embodiment of the present invention, first, in the block association step, the partial blocks corresponding to the respective parts of the human body are associated with the human imaging data imaged by the imaging means for imaging the person. Next, in the clothing data extraction step, clothing data that is data of the clothing portion is extracted from the partial blocks associated in the block association step. Furthermore, in the clothing shape specifying step, the clothing image pattern stored in the clothing image pattern storage means that stores the clothing data extracted in the clothing data extraction step and the clothing image pattern that is an image pattern for each type of clothing shape, and Are compared to determine the clothing image pattern that approximates the clothing data, and the shape of the person's clothing is specified based on the clothing image pattern. Then, in the apparel model image data creation step, based on the clothing of the shape identified in the clothing shape specifying step, creating a fashion model image data is data of a model image by modeling the clothing of a person. Then, in the clothing model image data output step, the clothing model image data created in the clothing model image data creation step is output. Thereby, one's own clothes can be reflected as clothes of a model image as they are. Furthermore, since the pattern patterns of model images are rich, a variety of model images can be created.

1 is a configuration diagram of a WEB providing system 1. FIG. It is a figure which shows the process until the avatar image reflecting the person's clothing is created. 3 is a block diagram showing an electrical configuration of a terminal device 3. FIG. 3 is a conceptual diagram showing various storage areas of a RAM 23. FIG. 3 is a conceptual diagram showing various storage areas of a flash memory 24. FIG. It is a flowchart of a clothing avatar creation process. It is a flowchart which shows the continuation of the flow of FIG. It is a flowchart of a person part extraction process. It is a flowchart of clothes cut-out processing. It is a flowchart of a clothes correction process. It is a flowchart which shows the continuation of FIG. It is a figure which shows an example of arrangement | positioning of a human body block. It is an example of the clothing template memorize | stored in the clothing template storage area 242. FIG. It is a figure which shows the brightness | luminance pattern of a plain shirt. It is a figure which shows the luminance pattern of the shirt of a grid | lattice pattern. It is a figure which shows the brightness | luminance pattern of the shirt which has a mark in the chest. It is a figure which shows the process which produces mark data based on a front point.

  Hereinafter, a terminal device 3 according to an embodiment of the present invention will be described with reference to the drawings. These drawings are used to explain technical features that can be adopted by the present invention. In other words, the configuration of the apparatus and the flowcharts of the various processes described are merely illustrative examples, and are not intended to be limited to only those unless specifically described. An avatar (corresponding to a “model image” of the present invention) is a character that appears on the screen as a substitute for a communication tool such as a homepage, a net game, or a chat.

  First, a WEB providing system 1 including a terminal device 3 as a configuration will be described with reference to FIG. The WEB providing system 1 includes a portable terminal device 3 that can be connected to a network 4 such as the Internet, a WEB server 5, and a terminal device 6 such as a stationary PC. As a WEB service, the WEB server 5 provides an avatar service linked to these services in addition to a normal communication service (such as an email service) and a net game service. The terminal devices 3 and 6 may be either a portable type or a stationary type, and are set as an example for convenience of explanation.

  The WEB server 5 provides a virtual space on the Internet, and provides an avatar that appears as a user's own substitute in the virtual space (see FIG. 2). The WEB server 5 stores a control unit (not shown) that controls an avatar and an object placed in the virtual space in response to an operation instruction from the terminal devices 3 and 6, and avatar information that associates the avatar with the object. An avatar information database (not shown). The WEB server 5 transmits user avatar information to the terminal devices 3 and 6 in response to processing requests from the terminal devices 3 and 6.

  The object arranged in the virtual space together with the avatar is an avatar item (display item) in the virtual space such as a hairstyle and clothing added to the avatar that is the user's own alternation. The user can attach and equip his avatar with the avatar item obtained in the virtual space from the terminal devices 3 and 6. An avatar item as an object is also arranged in the virtual space together with the avatar. The user can possess a plurality of avatar items in the virtual space.

  The actual processing for attaching the object to the avatar is performed by superimposing the object image on the avatar image in the virtual space, or synthesizing the avatar image and the object image. The virtual space may be a two-dimensional or three-dimensional space, and an avatar and an object may be displayed as a two-dimensional or three-dimensional display image in accordance with a two-dimensional or three-dimensional display on the avatar. In this embodiment, a three-dimensional avatar is assumed, but it may be two-dimensional.

  As illustrated in FIG. 1, the terminal device 3 includes a display 31 and a camera 32. The display 31 is provided with a touch panel 33 (see FIG. 3), and various operations can be performed by a touch operation on the display 31. The imaging data of the camera 32 is stored in the flash memory 24 (see FIG. 3) and can be displayed on the display 31. By connecting to the network 4, the terminal device 3 can use a homepage on the Internet, a net game, a chat, and the like, and can communicate with the server 5 and the terminal device 6. The terminal device 3 can also make a call with another terminal (not shown) through a portable telephone line.

  In the present embodiment, as illustrated in FIG. 2, the terminal device 3 extracts only the clothing from the user's whole body image data captured by the camera 32. And the terminal device 3 can create the clothing avatar data reflecting the color, pattern, mark, etc. of the extracted clothing. The clothing avatar data is image data of clothing (a shirt, pants, skirt, etc.) to be worn by the avatar. The created clothing avatar data is transmitted to the WEB server 5 and stored in the avatar information database as an avatar item. The clothing avatar data stored in the avatar information database can be acquired not only by the terminal device 3 but also by the terminal device 6 via the network 4. Thereby, in the terminal devices 3 and 6, the clothing avatar data acquired from the WEB server 5 can be reflected in the own avatar in the virtual space.

  Next, the electrical configuration of the terminal device 3 will be described with reference to FIG. The terminal device 3 is provided with a CPU 21 as a controller that controls the terminal device 3. The CPU 21 includes a ROM 22 that stores a BIOS, a main program, a RAM 23 that temporarily stores various data, a flash memory 24, a display 31, a camera 32, a touch panel 33, and a communication interface 35 (hereinafter, referred to as “CPU”). Communication I / F 35). The communication I / F 35 is for communicating with the network 4.

  Next, various storage areas of the RAM 23 will be described with reference to FIG. The RAM 23 includes a block arrangement storage area 231, a skin color reference value storage area 232, a luminance pattern storage area 233, a basic color storage area 234, a pattern data storage area 235, a mark data storage area 236, and a front point storage. At least an area 237 and a whole body image data storage area 238 are provided.

  The block arrangement storage area 231 stores the positions of human body blocks (corresponding to “partial blocks” of the present invention) A to F3 (see FIG. 12) arranged in whole-body image data described later. In the skin color reference value storage area 232, a color detected at the face position of human image data described later is stored as a skin color reference value. In the luminance pattern storage area 233, a luminance pattern which is a luminance distribution when analyzing the luminance of the fashion composition data described later is stored. In the basic color storage area 234, a color corresponding to the most frequent luminance in a luminance pattern described later is stored as a basic color. In the pattern data storage area 235, pattern data created based on a luminance pattern or the like is stored. In the mark data storage area 236, mark data created based on a luminance pattern or the like is stored. The front point storage area 237 stores the position of a front point (F), which will be described later, where the change in the luminance pattern is a predetermined value or more. Whole body image data captured by the camera 32 is stored in the whole body image data storage area 238.

  Next, various storage areas of the flash memory 24 will be described with reference to FIG. The flash memory 24 includes at least a program storage area 241, a clothing template storage area 242, a clothing avatar data storage area 243, and a 3D template storage area 244. The program storage area 241 stores a clothing avatar data creation program of the present invention, an editing program for editing the created clothing avatar data, and the like. The clothing template storage area 242 stores clothing templates for each category (shirt, trousers, skirt, etc.). The clothing avatar data created in the clothing avatar data creation process is stored in the clothing avatar data storage area 243. The 3D template storage area 244 stores a 3D template.

  Note that the 3D template includes front side clothing avatar data and back side clothing avatar data. In the server 5, 3D avatar information is created by combining the avatar data of the person and these two clothing avatar data. For example, human avatar data is composed of three-dimensional information (x, y, z). The clothing avatar data is composed of (X, Y) two-dimensional information. With the avatar data of the person facing the front and back, the clothing avatar data on the front side and the clothing avatar data on the back side are superimposed. Then, the z coordinate of the position where x = X and y = Y is defined as the Z coordinate of the clothing template. As a result, 3D avatar information having three-dimensional coordinates (X, Y, Z) is created. In the terminal devices 3 and 6 that have received the 3D avatar information, for example, the avatar can be rotated and displayed on the virtual space of the Internet.

  Next, the clothing template memorize | stored in the clothing template storage area 242 is demonstrated. In the clothing template storage area 242 of the flash memory 24, a plurality of clothing templates that are image data such as shirts and trousers are stored for each category. For example, as shown in FIG. 13, in the “shirt” category, clothing templates of shirts A to F having different shapes such as long sleeves, short sleeves, no collars, collars, and a turtleneck are stored. In the “trousers” category, clothing templates having long trousers, short trousers, knee-length trousers, boot cuts, buggy types, leggings, and other trousers A to F are stored. These templates are used to identify which category and type of clothing data the clothing data cut out from the whole body image of the person in the fashion avatar creation process described later.

  Next, the clothing avatar creation process executed by the CPU 21 will be described with reference to the flowcharts of FIGS. In this process, when the “clothing avatar creation application” is selected from the menu screen displayed on the display 31, the “clothing avatar creation program” of the present invention is read from the program storage area 241 of the flash memory 24 (see FIG. 5). It is issued and executed.

  As shown in FIG. 6, first, a photographing instruction message “Please shoot a whole body image” is displayed on the display 31 (S10). The user who sees this image the whole body of the person with the camera 32 of the terminal device 3 based on the instruction of the terminal device 3. And it is judged whether the shutter (not shown) of the terminal device 3 was pressed down (S11). While the shutter (not shown) is not pressed (S11: NO), the process returns to S11 and enters a standby state. When the shutter is pressed (S11: YES), the whole-body image data is displayed on the display 31, and confirmation buttons (not shown) for “YES” and “NO” are displayed on the lower side (S12).

  Next, it is determined whether or not “YES” has been selected (S13). If the user is not satisfied with the whole body image data displayed on the display 31 and “NO” is selected (S13: NO), the process returns to S10, and a shooting instruction message is displayed to prompt the user to take another shot. If “YES” is selected (S13: YES), it is stored in the whole body image data storage area 238 (see FIG. 4) of the RAM 23 (S14). At this time, whole body image data is displayed on the display 15. Next, a person part extraction process is executed for the whole body image data (S15).

  In this embodiment, in order to create 3D clothing avatar data, front and back whole body image data of a person is used. Therefore, when the two whole body image data are prepared, the person part extraction process is executed (S15). The following processing is also performed for each of the two whole body image data. For convenience of explanation, the flow of processing the front side whole body image data will be mainly described.

  Next, the person part extraction process will be described. As shown in FIG. 8, in the person part extraction process, first, a region that is supposed to be a part to be detected is extracted as a part candidate area from the whole body image using a region extraction method (S35). Next, the site | part of a site | part candidate region is pinpointed based on topology information, such as a connection relation and an adjacent relationship, with respect to the extracted site | part candidate area | region (S36). In this example, for example, the head, torso, upper right arm, right forearm, left upper arm, left forearm, right thigh, right lower leg, right foot, left thigh, left lower leg, A total of 12 sites on the left foot are identified.

  Then, as shown in FIG. 12, twelve rectangular human body blocks A to G3 are arranged for each identified part of the whole body image displayed on the display 31 (S37). For example, block A on the head, block B on the torso, block C1 on the upper right arm, block C2 on the right forearm, block D1 on the left upper arm, block D2 on the left forearm, block E on the waist, right thigh Block F1, the right lower leg block F2, the right leg block F3, the left thigh block G1, the left lower leg block G2, the left leg block G3 along the shape of each part of the human body Is done.

  Subsequently, returning to the flow of FIG. 6, a confirmation message “Is this the arrangement of the human body block?” And confirmation buttons (not shown) of “YES” and “NO” are displayed below the display 31. (S16). The user who sees this looks at the arrangement of the human body block displayed on the display 31 and confirms whether or not the human body block is arranged at an appropriate position. Then, it is determined whether or not “YES” is selected (S17).

  Here, when there is a human body block whose position is shifted and “NO” is selected by the user (S17: NO), “manual mode” is set (S18). In the manual mode, the position of the human body block can be adjusted manually. As shown in FIG. 12, the user moves the human body block displayed on the display 31 while touching it with a finger. The movement of the finger is detected by the touch panel 33 (see FIG. 3), and the position of the human body block is changed according to the detection result. For example, when the position of the block C1 is shifted from the upper right arm, the position and angle of the block C1 are adjusted while bringing the finger into contact with the block C1. As a result, the block C1 can be accurately aligned with the position of the upper right arm.

  On the lower side of the display 31, an OK button for determining the arrangement of the human body block is displayed. Next, it is determined whether or not “OK” is selected (S19). Until the determination button is selected (S19: NO), the process returns to S18 and the manual mode is maintained. As described above, since the arrangement of the human body blocks automatically arranged by the CPU 21 can be manually adjusted, even if there is a deviation in the automatic arrangement by the CPU 21, it can be adjusted to an accurate position.

  When the determination button is selected (S19: YES), these block arrangements are stored in the block arrangement storage area 231 (see FIG. 4) of the RAM 23, the edges around the human body block are detected, and the whole body image data The person image data from which the background has been removed is extracted (S20). If the arrangement of the human body blocks displayed on the display 31 is good and “YES” is selected (S17: YES), the process proceeds to S20, and person image data is extracted in the same manner as described above (S20). Further, the extracted person image data and confirmation buttons “OK” and “NO” are displayed on the display 31 (S21). Next, it is determined whether or not an “OK” confirmation button has been selected (S22).

  For example, when the background is not sufficiently removed and the edge of the person is unclear, the user selects the confirmation button “NO”. In this case (S22: NO), the process returns to S18 and the manual mode is reset. By causing the user to adjust the position of the human body block and to execute the process of extracting the person image data again, person image data with a clear edge can be created. If “OK” is selected (S22: YES), since only the person can be extracted from the whole body image data, the process proceeds to the flow of FIG. 7, and “clothes” for cutting out only the clothing based on the extracted person image data is displayed. The “cutout process” is executed (S25).

  Next, the clothing cutout process will be described. As shown in FIG. 9, first, the face position is recognized for the person image data (S41). As the face recognition method, various conventionally known methods can be used. For example, an algorithm such as a Harris operator or SIFT (Scale Invariant Feature Transform) can be used to extract facial feature points. In this embodiment, by detecting feature points of eyes, nose and mouth from block A, the position of the face on the head is recognized. Next, it is determined whether or not the position of the face has been correctly recognized (S42).

  If the face position is recognized (S42: YES), the color is extracted from the recognized face position (S43). Specifically, the color is extracted from the area formed by the eyes, nose, and mouth where the feature points are detected. The face is a part that is exposed to the outside and is the most standard part that represents the skin color of the person. Further, by extracting the color from the region formed by the eyes, nose, and mouth, the skin color is accurately extracted regardless of the color of the hair. The color of such a part is stored in the skin color reference value storage area 232 (see FIG. 4) of the RAM 23 as a skin color reference value (S45).

  On the other hand, if the feature points of the eyes, nose, and mouth are not accurately detected from the block A and the face position is not correctly recognized (S42: NO), the color extraction position receiving process is executed (S44). . In this reception process, the user is caused to specify the position to be used as the skin color reference. The display 31 displays a skin color reference mark (not shown) and an OK button. The skin color reference mark can be moved on the touch panel 33. When the OK button is selected by the user, the color at the position of the flesh color reference mark is stored as the flesh color reference value in the flesh color reference value storage area 232 (see FIG. 4) of the RAM 23 (S45).

  Next, for each human body block B to G3 (see FIG. 12) excluding the head block A, pixel color data in each block is acquired one by one (S46). Further, by comparing the acquired pixel color data with the skin color reference value stored in the skin color reference value storage area 232, it is determined whether or not the acquired pixel color data is a skin color (S47). If the color data of the pixel is a skin color (S47: YES), the pixel image is deleted from the person image data because it is a region that is not clothing in the person image data (S48).

  On the other hand, if the color data of the pixel is not skin color (S47: NO), it is a clothing area, so it is not deleted from the human body image data, and it is determined whether or not comparison has been completed for all pixels. (S49). If all the pixels have not been completed (S49: NO), the next pixel is acquired (S51), and the process returns to S46 to perform the same processing (S47, S48). The process is repeated until the comparison is completed for all the pixels (S49: NO). When the comparison is completed for all the pixels (S49: YES), the skin color area is deleted from the person image data. The area that remains without being deleted becomes “clothing data”.

  Next, based on the block arrangement stored in the block arrangement storage area 231 of the RAM 23, the clothing data is divided into an upper body part and a lower body part (S50). In the example shown in FIG. 12, the portions corresponding to the blocks B, C1, C2, D1, and D2 are the upper body portions, the portions corresponding to the blocks E, F1 to 3, and G1 to 3 are the lower body portions in the two regions. Divided. Thereby, it can divide into the clothing data of a shirt, and the clothing data of trousers (or skirt etc.), respectively. For example, when only the clothing avatar data of a shirt is created, the portions corresponding to the blocks E, F1 to 3 and G1 to 3 which are lower body portions may be deleted. Then, returning to the flow of FIG. 7, clothing correction processing is executed (S26).

  The clothing correction process will be described. In the present embodiment, only the clothing data of the shirt will be described for convenience of explanation. As shown in FIG. 10, first, a binarization process of clothing data is executed (S52). This is to clarify the shape of clothing. Next, the shape of the clothing template stored in the clothing template storage area 242 (see FIG. 5) of the flash memory 24 is changed along with the clothing data in accordance with the arrangement of the human body blocks arranged in the person image data (S53). ). In the example shown in FIG. 13, the left sleeve of the shirt of the clothing data is raised upward. Therefore, the change which raises the position of the left sleeve of the clothing template of all the shirts is performed. Then, a matching process between the template whose shape has been changed and the clothing data is executed (S54). Thereby, even if the posture of the person imaged by the camera 32 is not upright, the shape of the clothing template can be changed according to the posture of the person, so that the matching process can be performed accurately.

  In the example shown in FIG. 13, since the clothing template having the highest correlation with the clothing data is the shirt D, an area along the shape of the shirt D is cut out from the person image data obtained by removing the background from the whole body image data in S20 ( S55). Further, the person image data of the cut-out area is synthesized with the clothing template of the shirt D (S56). In this manner, the clothing composition data of the shirt portion is extracted from the person image data. Then, on the display 31, clothing composition data as an extraction result and confirmation buttons (not shown) of “YES” and “NO” are displayed (S57).

  Next, it is determined whether or not “YES” has been selected (S58). When it is determined that “YES” has been selected (S58: YES), a postalization process for the clothing composition data is executed (S59). In the posterization process, the graphic gradation is changed. A normal full color graphic is expressed with 256 gradations for each of RGB. By changing the gradation, for example, an unclear portion where a plurality of colors are intertwined can be expressed with a small number of colors, so that the color arrangement can be clarified. That is, the color and pattern of the shirt can be clarified.

  On the other hand, when the user is not satisfied with the clothing composition data displayed on the display 31 and it is determined that “NO” has been selected (S58: NO), the composition processing with the clothing template having the second and third highest correlations Is executed (S62). Further, the two synthesized clothing composition data and the confirmation button are displayed on the display 31 (S63). The user touches the user with his / her favorite one and selects the “OK” confirmation button. Next, it is determined whether or not the clothing composition data has been selected (S64). When “OK” is selected and the clothing composition data is selected (S64: YES), the posterization process is executed (S59).

  If the user cannot select the clothing composition data and it is determined that “NO” is selected (S64: NO), the process returns to S53, the matching process with the clothing template is executed again, and the clothing composition data Extraction is performed (S54 to S57).

  Next, the luminance of the clothing composition data is analyzed (S60). Here, the clothing composition data is scanned laterally at predetermined intervals in order from the top, and the change in luminance is measured. The luminance varies depending on the color, pattern, mark, etc. of the clothing. For example, since the brightness pattern of the plain shirt shown in FIG. 14 is only one color, the brightness does not change and becomes a straight line. The luminance pattern of the shirt having a lattice pattern shown in FIG. 15 is colored and colorless, so that the luminance changes up and down at every predetermined interval, and the tooth shape has a wave shape. The luminance pattern of the shirt having a circular one-point mark on the chest shown in FIG. 16 has a shape in which the center is recessed because the luminance is rapidly reduced at a position corresponding to the mark.

  The luminance pattern analyzed in this way is stored in the luminance pattern storage area 233 (see FIG. 4) of the RAM 23, and the point where the change in the luminance pattern is a predetermined value or more indicates the front point ( F) are stored in the front point storage area 237 (see FIG. 4) of the RAM 23 (S61).

  Next, proceeding to the flow of FIG. 11, it is determined whether or not the luminance pattern is not constant for the luminance pattern stored in the luminance pattern storage area 233 of the RAM 23 (S72). When the brightness pattern is not constant as in the brightness patterns shown in FIGS. 15 and 16 (S72: YES), the most frequently used brightness is extracted as the basic color of the clothing (shirt in this embodiment) (S73). ). The most frequently used luminance portion is a color that has the largest proportion of the overall composition data. By making such a color the basic color of a clothing avatar, the basic color of the actual clothing of a person can be reflected. The extracted basic colors are stored in the basic color storage area 234 (see FIG. 4) of the RAM 23.

  Next, it is determined whether or not the luminance pattern has a period (S74). For example, when there is a period as in the luminance pattern shown in FIG. 15 (S74: YES), the pattern is based on the period and the position of the front point stored in the front point storage area 237 (see FIG. 4) of the RAM 23. Data is created. This pattern data is stored in the pattern data storage area 235 (see FIG. 4) of the RAM 23 (S75). Thereafter, the process proceeds to S76.

  On the other hand, when there is no period in the luminance pattern (S74: NO), it is subsequently determined whether or not the luminance pattern is locally changed (S76). For example, in the case of a local change as in the luminance pattern shown in FIG. 16 (S76: YES), the luminance of the locally changed portion and the front point storage area 237 of the RAM 23 (see FIG. 4) Mark data is created based on the stored position of the front point. For example, as shown in FIG. 17, when a one-point mark on the chest of a shirt is scanned, a plurality of front points are detected and stored. If only the front points are extracted, the front points are arranged along the boundary of the mark.

  Therefore, by connecting a front point and a front point, which are close to each other and have the same brightness of the clothing data at that position, by a Bezier curve, a circular mark can be expressed in a shape close to the real object. Here, when the front point is scanned in only one direction, it may be difficult to detect a mark boundary that is substantially parallel to the scanning line. Therefore, the front point may be detected in a lattice shape by setting the scanning direction to two directions, up and down and left and right. By detecting the front points in a lattice shape, the distance of proximity can be limited to within the diagonal distance of the lattice. The mark data created in this way is stored in the mark data storage area 236 (see FIG. 4) of the RAM 23 (S77). If the luminance pattern has not changed locally (S76: NO), it is not a mark, and the process directly proceeds to S78.

  If the luminance pattern has a period and there is a part that changes locally (S74: YES, S76: YES), pattern data and mark data are created, and the pattern data storage area 235 of the RAM 23 and the mark The data is stored in the data storage area 236 (S75, S77). Further, when the luminance pattern is constant as shown in FIG. 14 (S72: NO), an average luminance value is calculated, and the calculated value is extracted as a basic color of clothing, and a basic color storage area 234 of the RAM 23 is obtained. (See FIG. 4) (S82).

  Subsequently, the pattern composition data stored in the pattern data storage area 235 of the RAM 23, the mark data stored in the mark data storage area 236, and the basic color storage area 234 are stored for the clothing composition data created in S 65. The basic colors are combined to create clothing avatar data (S78).

  Subsequently, the created clothing avatar data is displayed on the display 31, and a confirmation message as to whether or not to edit the clothing avatar data and confirmation buttons “YES” and “NO” are displayed below the display 31. (S79). Then, it is determined whether or not the “YES” confirmation button has been selected (S80). When “YES” is selected (S80: YES), the editing application is activated (S81), and the editing screen is displayed. On the editing screen, enlargement / reduction, rotation, drawing, color scheme, and the like of the clothing avatar data can be selected, and the clothing avatar data is edited according to the selected process. Therefore, the clothing avatar data can be freely changed according to the user's preference. It is also possible to correct unclear portions of the clothing avatar data.

  On the other hand, when the “NO” confirmation button is selected (S80: NO), the flow returns to the flow of FIG. Note that the processing up to this point is similarly executed for the whole body image on the back of the person in addition to the whole body image on the front of the person. Accordingly, as the clothing avatar data, front data and back data are created. Next, these clothing avatar data are pasted on the 3D template stored in the 3D template storage area 244 of the flash memory 24 (S27).

  Subsequently, a confirmation message “Do you want to save the clothing avatar data?” And confirmation buttons “YES” and “NO” are displayed on the display 31, and the clothing avatar data is displayed based on the selection of the confirmation button by the user. It is determined whether or not to save (S28). When “YES” is selected (S28: YES), the created clothing avatar data is stored in the clothing avatar data storage area 243 of the flash memory 24 (S29). When “NO” is selected (S28: NO), a confirmation message “Do you want to send the clothing avatar data to the server?” And confirmation buttons “YES” and “NO” are displayed on the display 31; Based on the selection of the user's confirmation button, it is determined whether or not to transmit the clothing avatar data (S30).

  If “YES” is selected (S30: YES), the clothing avatar data is transmitted to the server 5 via the network 4 (S31), and the process is terminated. If the NO confirmation button is selected (S30: NO), the process ends. In the server 5, the received clothing avatar data is stored as an avatar item as an object. And according to the request | requirement of the terminal devices 3 and 6, the avatar information which linked | related the avatar and the object is transmitted to the terminal devices 3 and 6. FIG. Therefore, for example, by sending the clothing avatar data created by the terminal device 3 to the server 5 and storing it in the server 5, by accessing the server 5 from the terminal device 6, the clothing avatar data is obtained in the virtual space of the Internet. The reflected avatar can appear.

  In the above description, the camera 32 shown in FIGS. 1 and 3 corresponds to the “imaging means” of the present invention, and the clothing template storage area 242 of the flash memory 24 shown in FIG. Is equivalent to. 6 corresponds to the “person part extracting process” of the present invention, and the CPU 21 executing the process of S25 of FIG. 7 corresponds to the “clothes cutting process” of the present invention. The CPU 21 that executes the process of S54 of FIG. 10 corresponds to the “clothing shape specifying means” of the present invention, the CPU 21 that executes the process of S78 of FIG. 11 corresponds to the “clothing model image data generating means” of the present invention, and S7 The CPU 21 that executes the process of S31 corresponds to the “fashion model image data output means” of the present invention.

  Further, the CPU 21 that executes the processes of S60, 61, 72, 73, 74, 75, 76, and 77 in FIGS. 10 and 11 corresponds to the “design extracting means” of the present invention. Further, the CPU 21 that executes the process of S46 of FIG. 9 corresponds to the “skin color area specifying unit” of the present invention. Further, the CPU 21 that executes the process of S62 in FIG. 1 corresponds to the “shape changing means” of the present invention. Further, the CPU 21 that executes the process of S80 in FIG. 11 corresponds to the “fashion model image data changing unit” of the present invention. Further, the CPU 21 that executes the process of S18 in FIG. 6 corresponds to the “block position changing unit” of the present invention.

  As described above, the terminal device 3 according to the present embodiment can create the clothing avatar data that models only the clothing of the person based on the whole body image data of the person imaged by the camera 32. In the clothing avatar creation process by the CPU 21, a person part extraction process for extracting a person part from the whole body image data is executed. Next, based on the extracted person part, person image data is extracted from the whole body image data. Furthermore, a clothing cutout process is performed in which the skin color region is deleted from the person image data and the clothing portion is cut out. Next, in order to identify the category of clothing, matching processing with the plurality of clothing templates stored in the flash memory 24 is performed on the clothing data created by the clothing cutout processing.

  Then, the shape of the category of the template having the highest correlation is cut out from the person image data, and is combined with the template to create the clothing combined data. Subsequently, with respect to the clothing composition data, the luminance pattern and the front point whose luminance change is equal to or greater than a predetermined value are analyzed, and clothing avatar data is created based on the analysis information. The clothing avatar data is transmitted to the server 5 by the user's selection.

  In the server 5, the transmitted clothing avatar data is stored as an avatar item as an object. And according to the request | requirement of the terminal devices 3 and 6, the avatar information which linked | related the avatar and the object is transmitted to the terminal devices 3 and 6. FIG. Thereby, the avatar which reflected the clothing avatar data can be made to appear in the virtual space of the Internet. Furthermore, since there are abundant patterns of avatar clothing, it is possible to create a wide variety of avatar images and to further improve the function of the avatar as a communication tool.

  The present invention is not limited to the above-described embodiment, and various modifications can be made. For example, in the above embodiment, in order to create the 3D clothing avatar data, the camera 32 captures the whole body image data about the front and back of the person, but creates the clothing avatar data for only the front (back only). In this case, the whole body image data may be only the front or back side.

  Further, for example, 3D clothing avatar data may be created using only the whole body image data on the front side of a person. When only the front-side whole body image data is used, for example, front-side whole body image data excluding the mark data (see S77 in FIG. 11) is used instead of the back-side whole body image data.

  In the above embodiment, shirts, trousers, skirts, and the like are given as examples of clothing categories, but clothing avatar data may also be created for shoes, gloves, and decorative items (necklaces, bracelets, etc.). Good.

  Moreover, although the terminal device 3 was demonstrated as a portable terminal in the said embodiment, stationary type and a notebook type PC may be sufficient and the shape of a terminal is not limited.

  Moreover, in the said embodiment, although the clothing avatar data was produced with the terminal device 3 and the web provision system 1 transmitted to the server 5 was demonstrated, you may make it produce the clothing avatar data with the server 5, for example. In this case, whole body image data captured by the terminal device 3 may be transmitted to the server 5, and the clothing avatar data may be created by the server 5 based on the whole body image data as in the above embodiment.

  In the embodiment described above, the CPU 21 that transmits the clothing avatar data to the server 5 has been described as equivalent to the “clothing model image data output unit” of the present invention. For example, the CPU 21 that displays the clothing avatar data on the display 31 It may be.

3 Terminal device 21 CPU
23 RAM
24 Flash memory 31 Display 32 Camera 231 Block arrangement storage area 232 Skin color reference value storage area 233 Luminance pattern storage area 234 Basic color storage area 235 Pattern data storage area 236 Mark data storage area 237 Front point storage area 241 Program storage area 242 Clothing template Storage area 243 Clothing avatar data storage area

Claims (8)

A model image creation device that creates clothing model image data that is an image modeling the shape of a person's clothing,
A clothing image pattern storage means for storing a clothing image pattern which is an image pattern for each type of shape of a person's clothing;
Block associating means for associating partial blocks corresponding to the respective parts of the human body with the image data of the person imaged by the imaging means for imaging the person;
Clothing data extraction means for extracting clothing data that is data of a clothing part from the partial block associated by the block association means;
By comparing the clothing data extracted by the clothing data extracting means and the clothing image pattern stored in the clothing image pattern storage means, the clothing image pattern that approximates the clothing data is determined, Based on a clothing image pattern, clothing shape specifying means for specifying the shape of the person's clothing,
Clothing model image data creating means for creating clothing model image data that is model image data of the person's clothing based on the shape of the clothing identified by the clothing shape identifying means;
A model image creation apparatus comprising: a clothing model image data output unit that outputs the clothing model image data created by the clothing model image data creation unit. In the image data of the person, comprising a design extracting means for extracting a design including a clothing pattern from a part corresponding to the shape of the clothing specified by the clothing shape specifying means,
The clothing model image data creation means includes:
The model image according to claim 1, wherein the clothing model image data is created by synthesizing the shape of the clothing specified by the clothing shape specifying means and the design extracted by the design extraction means. Creation device. The clothing data extraction means includes:
Within the partial block associated by the block association means, comprises a skin color area specifying means for specifying a skin color area,
The model image creating apparatus according to claim 1, wherein an area obtained by removing the skin color area specified by the skin color area specifying unit from the partial block is extracted as the clothing data.   The skin color area specifying unit acquires a reference value for specifying the skin color area from a partial block corresponding to the face portion of the person among the partial blocks associated with the imaging data by the block association unit. The model image creating apparatus according to claim 3, wherein the skin color region is specified based on the reference value. The clothing shape specifying means is:
A shape changing unit for changing the shape of the clothing image pattern stored in the clothing image pattern storage unit along the shape of the clothing data extracted by the clothing data extracting unit;
5. The model image according to claim 1, wherein the clothing data extracted by the clothing data extracting unit is compared with the clothing image pattern whose shape has been changed by the shape changing unit. Creation device.   6. A clothing model image data changing unit for changing a shape and design of a clothing model image corresponding to the clothing model image data created by the clothing model image data creating unit. A model image creation device according to claim 1.   The model image creating apparatus according to claim 1, further comprising a block position changing unit that changes a position of the partial block associated by the block associating unit. A model image creation program of a model image creation device that causes a model image creation device to create clothing model image data that is an image obtained by modeling the shape of a person's clothing,
A block associating step for associating each of the partial blocks corresponding to each part of the human body with the imaging data of the person imaged by the imaging means for imaging the person;
A clothing data extraction step of extracting clothing data that is data of a clothing part from the partial block associated in the block association step;
By comparing the clothing data extracted in the clothing data extraction step with the clothing image pattern stored in the clothing image pattern storage means for storing the clothing image pattern that is an image pattern for each type of clothing shape Determining the clothing image pattern that approximates the clothing data, and specifying the shape of the person's clothing based on the clothing image pattern;
A clothing model image data creation step of creating a fashion model image data is data of the clothing shape based on the clothing of the shape specified in the specifying step, the model image by modeling the clothing of the person,
A model image creation program for causing a model image creation device to execute a clothing model image data output step for outputting the clothing model image data created in the clothing model image data creation step.

Images