JP2003269936A - Automatic size-measuring method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To automatically measure a size by photographing a subject with his or her clothes on. <P>SOLUTION: The subject gets on a rotary table 10 by winding a belt around a size-measuring position and by tightly fitting the clothes to his or her body. The subject is photographed with a camera 12, and the obtained data are fed to a computer 14. Three-dimensional shape data of the subject is generated from the inputted image data in the computer 14, and the size of a desired position is measured from the three-dimensional shape data. Instead of the belt, the subject may wear an accessary formed of a non-mirror surface reflecting material or may wear an orthosis with a specific marker or a random pattern formed, and thereby the three-dimensional shape data can be also created by a stereo method as well. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic measuring method, and more particularly to a technique for automatically measuring while wearing clothes.

[0002]

2. Description of the Related Art Conventionally, a technique for measuring the shape of a human body while wearing clothes has been proposed. For example, in Japanese Unexamined Patent Publication No. 5-203434, a subject wearing clothes is irradiated with continuous sound waves, and the influence of clothes is eliminated by utilizing the fact that reflected sound waves from clothes interfere with reflected sound waves from the human body.
A technique for acquiring only information from the human body and measuring the human body shape is described. Specifically, 20 kHz to 40 kHz
The ultrasonic wave of the human body is reflected, the reflected sound wave from the human body surface and the reflected sound wave from the clothing surface are acquired, and the reflected wave signal of the human body is determined by removing the clothes reflected sound wave component determined by the sound wave reflectance of the clothes from the received signal. Ask for.

However, it is difficult to measure the human body shape with high accuracy because sound waves have poor directivity as compared with light and the like, and are greatly affected by diffraction and the like. In addition, soundproofing measures such as wrapping sound wave transmitters and receivers with sound absorbing material and attaching sound absorbing material to walls, ceilings, floors, etc. are required.
There is also a problem that the facility is large and expensive. Furthermore, some subjects may have a resistance to the irradiation of ultrasonic waves on their body. Therefore, in general, it is preferable to use light or the like to measure the human body shape of the subject in a clothes state, that is, it is preferable to measure the human body shape from an image taken by a camera or the like.

Japanese Patent Publication No. 7-76690 describes a technique for measuring a human body shape from image data obtained by photographing a subject. In this conventional technique, an infrared camera captures an image of a subject in a dressed state and acquires contour data of the human body. Regarding the contour of the obtained image data, the exposed portion such as the hand or face is the contour, and the portion covered by the clothing is the contour of the clothing. Three-dimensional image data is synthesized from the contour data. Next, temperature data corresponding to each coordinate of the three-dimensional image is acquired, a thermograph is created by color-coding the temperature of each coordinate, and image data corresponding to the contour of the naked human body is created from this thermograph. Creating.

[0005]

However, in this prior art, a highly accurate thermograph is indispensable for creating contour data of a naked human body. Specifically, we have created a thermograph based on the data about body temperature and physical condition of the day that we asked from the subject, and factors that affect body temperature such as gender and age, and it is complicated to acquire these data. However, there is also a problem that the burden on the subject increases because it is necessary for the subject to accurately convey his / her physical condition data.

[0006] It is desirable for the subject to be able to measure the clothes as they are, but it is troublesome to convey his / her own body temperature and physical condition data instead of wearing clothes, and this kind of personal information should be given to a third party. It is expected that there will be some resistance to transmitting data.

The present invention has been made in view of the above-mentioned problems of the prior art, and the object thereof is to measure the size of a subject in a clothes state without needing to obtain body temperature data of the subject and the psychological burden on the subject. It is to provide an automatic measuring method that can greatly reduce the

[0008]

In order to achieve the above object, the present invention is a method for measuring a subject in a clothes state, which is a step of photographing a subject in which a belt is wrapped around the clothes at a measurement site. And a step of creating three-dimensional shape data of the measurement site from a plurality of images obtained by photographing, and a step of obtaining a size of the measurement site based on the three-dimensional shape data. In the step of creating three-dimensional shape data, the existence probability of the subject in a predetermined voxel space is calculated by using the contour data of the subject in a plurality of images obtained by photographing the subject from a plurality of directions. It is preferable to create the three-dimensional shape data.

Further, the present invention is a method for measuring a subject in a clothing state, which comprises the steps of photographing a subject wearing a brace made of a non-specular reflection material at the measuring portion, and a plurality of images obtained by photographing. From the step of creating three-dimensional shape data of the measurement site and the step of obtaining the size of the measurement site based on the three-dimensional shape data. In the step of creating the three-dimensional shape data, it is preferable to create the three-dimensional shape data based on the parallax (stereo difference) of two images obtained by taking images at two imaging positions that are separated from each other by a predetermined distance. Therefore, it is preferable to calculate the parallax by using points on the isoluminance line in the image as corresponding points.

Further, the present invention is a method for measuring a subject in a clothes state, which comprises the steps of photographing a subject wearing a brace having a specific mark on a measuring portion, and a plurality of images obtained by photographing. The method is characterized by including a step of creating three-dimensional shape data of the measurement site and a step of obtaining a size of the measurement site based on the three-dimensional shape data. In the step of creating the three-dimensional shape data, it is preferable that the three-dimensional shape data is created based on the parallax of two images obtained by photographing at two photographing positions separated from each other by a predetermined distance. It is preferable to calculate using the marks as corresponding points.

Further, the present invention is a method for measuring a subject in a clothes state, which comprises a step of photographing a subject wearing a brace having a random pattern on a measuring portion, and a plurality of images obtained by photographing. The method is characterized by including a step of creating three-dimensional shape data of the measurement site and a step of obtaining a size of the measurement site based on the three-dimensional shape data. In the step of creating the three-dimensional shape data, it is preferable that the three-dimensional shape data is created based on the parallax of two images obtained by photographing at two photographing positions separated from each other by a predetermined distance. It is preferable to calculate using a random pattern as the corresponding portion.

As described above, according to the present invention, the subject can automatically measure the size by wearing a predetermined belt or a brace made of a predetermined material, so that the psychological burden on the subject is reduced. The predetermined belt has a function of bringing the clothes into close contact with the human body, and the contour line obtained from the image should be almost naked. To create three-dimensional shape data from image data, a silhouette method for calculating the existence probability in a predetermined voxel space or a stereo method using the parallax of two cameras can be used. In the stereo method, it is necessary to extract corresponding points between two images, but it is possible to facilitate the search for corresponding points by having the subject wear a specific orthosis. By wearing a brace made of non-specular reflection material (or perfect diffuse reflection material), the light from the light source is diffusely reflected by the brace, so the brightness of the image obtained by the camera is determined by the physical distance between the camera and the brace. Therefore, equal-luminance lines are formed in the left and right images, respectively. By extracting the isoluminance line t of the right image corresponding to the isoluminance line t in the left image and extracting the corresponding points on the corresponding isoluminance line,
3 even by the stereo method even in the part without the edge
It is possible to generate a three-dimensional shape and ensure the accuracy of the three-dimensional shape.
In the case of a device having a mark, the mark can be directly used as a corresponding point in the stereo method.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

<First Embodiment> FIG. 1 shows the configuration of an automatic measuring system according to the present embodiment. The automatic measuring system includes a rotary table 10, a camera 12 and a computer 14.

The turntable 10 is composed of a fixed portion 10a and a turntable 10b, and a subject rides on the turntable 10b in a clothes state. Since the rotary table 10 is rotationally driven by a rotary mechanism (not shown), a holder for the subject to secure balance can be provided on the rotary table 10b.

The camera 12 is, for example, a CCD camera,
The rotary table 10 is installed at a predetermined relative position. The camera 12 may be a visible light camera or an infrared camera. When the turntable 10 accelerates to a predetermined rotation speed and rotates at a constant speed, the camera 12 captures an image of a subject in a clothing state and acquires image data. Camera 12
Takes the subject in all directions, for example, in 3 degree increments of 10 degrees.
Acquire image data of 6 images. The imaging range may be the whole body of the subject, or if the measurement site is a bust, waist, hip, etc., it may be only the site including these.
The camera 12 and the computer 14 are connected by wire or wirelessly, and the image data obtained by the camera 12 is supplied to the computer 14. Image data obtained by the camera 12 is stored in a flash memory or other storage medium,
It is also possible to connect this storage medium to the computer 14 and supply the image data to the computer 14.

A block diagram of the computer 14 is shown in FIG. The computer 14 has a camera 12
Interface I / F 14 for inputting image data from
a, CPU 14b, image storage unit 14c, shape storage unit 14
d and a ROM 14e for storing the processing program. The image data captured from the interface I / F 14a is stored in the image storage unit 14c. CPU
14b reads the image data stored in the image storage unit 14c according to the processing program stored in the ROM, generates three-dimensional shape data of the subject by the silhouette method described later, and stores the three-dimensional shape data in the shape storage unit 14d. Then, using the obtained three-dimensional shape data, the size of the desired measurement site, for example, the bust, waist, hips, etc., is measured and the CRT is measured.
Or display on a display device such as a liquid crystal display or output to a printing device such as a printer.

In the present embodiment, the size of the desired part of the subject is automatically measured simply by placing the subject on the turntable 10 while wearing the clothes. However, the contour of the image data obtained by the camera 12 is the contour of the clothes worn by the subject, and does not necessarily match the contour of the human body (contour in the naked state). Therefore, in the present embodiment, the subject wears a predetermined belt on the measurement site to eliminate the influence of clothes as much as possible.

FIG. 3 shows an example of a belt worn by a subject. The belt 16 has a body portion 1 having a predetermined width.
6b and Velcro (registered trademark) 16a at both ends. For example, when measuring a bust, the subject winds the belt 16 on the bust from above the clothes and fastens it with Velcro (registered trademark) 16a. Belt 1
6 is for eliminating an unnecessary space between the clothing and the human body as much as possible, and the clothing may be worn with a tension that allows the clothing to be in close contact with the human body. Usually, when measuring with a measure or the like, it is necessary for a specialized operator who is familiar with the measurement to operate the measure in consideration of reproducibility and accuracy, but the belt 16 in the present embodiment makes the clothing and the human body adhere to each other appropriately. Therefore, it is not necessary to have skill like major operation,
Can be worn by the subject himself.

In FIG. 4, the belt 16 is attached to the measurement site of the subject.
A model diagram of a state in which is wrapped is shown. In addition, in FIG. 4, a human body and clothes are schematically illustrated in a cylindrical shape.
(A) is a model diagram of a state in which the belt 16 is not wound, and the contour 18a of the clothing is the contour 18b of the human body itself.
Is hidden. (B) is a model diagram in the case where the belt 16 is wrapped around the measuring portion, and the contour 18b of the human body appears at the portion where the belt 16 makes the clothing adhere to the human body. The camera 12 acquires an omnidirectional image of the portion where the contour 18b of the human body appears, and the computer 14 generates three-dimensional shape data of the portion. This allows
Accurate measurement of the human body is possible without the influence of clothes. The three-dimensional shape data itself may be the entire human body. The three-dimensional shape data of the portion around which the belt is wound indicates the accurate contour of the human body (contour in the naked state), and the other three-dimensional shape data indicates the contour of the clothing. Belt 1 if you need to measure your bust, waist, or hips
Wrap 6 around these parts to get images in a batch,
It is preferable to collectively generate the three-dimensional shape data.

The processing of the computer 14 for generating three-dimensional shape data from the obtained image data will be described below.

FIG. 5 shows an image processing flowchart in the computer 14. First, the computer 14 performs calibration for adjusting the relative position between the camera 12 and the turntable 10 (S10).
1). In this calibration process, for example, as a relative positional relationship between the turntable 10 and the camera 12,
For example, the position (x0, y0, z0) of the camera 12 using a coordinate system (xyz coordinate system) with the rotary table 10 as a reference.
And the angles α, β, γ formed by the optical axes of the camera 12 can be used. Here, α is the rotation angle of the camera 12 about the optical axis, β is the angle formed by the optical axis and the y axis when the camera 12 is orthographically projected on the yz plane, and γ is the angle when the camera 12 is orthographically projected on the xy plane. It is the angle between the optical axis and the y-axis. By making the optical axis of the camera 12 substantially parallel to the rotation plane of the turntable 10,
The angle parameter can be simplified by setting α = γ = 0, and in this case, the angle data indicates the depression angle of the camera 12 with respect to the rotary table 10. In the calibration, the internal parameter (perspective ratio) of the camera 12 may be determined together. The perspective ratio is calculated by T / L from the distance L between the camera 12 and the reference object and the height T of the reference object by photographing the reference object on the rotary table 10. The enlargement / reduction ratio of the size of the object projected on the screen is determined by this perspective ratio.

After the relative positional relationship between the turntable 10 and the camera 12 is determined, the camera 12 takes an image of a target object including the background (a subject wearing the belt 16) (S10).
2). The background is preferably a color (for example, blue) different from that of the subject's clothes and belt. The reason is that only the contour (silhouette) of the subject can be easily extracted from the image data by the known chroma key technique.

Omnidirectional image data of the subject (36 images)
After acquiring, the silhouette image of the subject is created from the obtained image data (S103). This processing can use the chroma key technique as described above, and a silhouette image is created by excluding a known background image and rotation table image from the image data in advance. After the silhouette image is created, the process proceeds to the voting process (S104) and the polygon creation process (S105) for creating the three-dimensional shape data of the subject from the silhouette image. These will be described below.

In the voting process, a three-dimensional voxel space consisting of a plurality of voxels is first assumed. The voxel space can have, for example, a cylindrical shape. A cylinder in voxel space is cut along a plurality of planes perpendicular to the central axis, and cut along a plurality of planes including the central axis and parallel to the central axis. Further, the cylinder is cut by a plurality of concentric curved surfaces having the central axis as an axis. Each element of the cylinder obtained by cutting the cylinder as described above is a voxel. For each voxel in the voxel space, voting (voting) processing is performed using the plurality of (for example, 36) silhouette images (B1 to B36) obtained in S103.

FIG. 6 schematically shows the voting process. It should be noted that FIGS. 10A and 10B show a plurality of image data A1 and A obtained by photographing the target object 23, respectively.
2, ... AN and silhouette images B1, B2, ..., BN obtained from these plural images are shown.
A giraffe toy is shown as the target object 23, and the silhouette image also shows the outline of this giraffe. However, in the present embodiment, the target object is the subject wearing the belt 16, and the silhouette image also provides the outline of the human body. It's easy to understand. For the obtained silhouette image BN,
With the projection center of the camera 12 as the apex, the silhouette image BN
A cone-shaped region having a cross-sectional shape of the inner silhouette image is generated. If this area is assumed to be the assumed existence area, the subject will always exist inside this assumed existence area. In the voxel space, "1" is voted for all voxels corresponding to this hypothetical existing area. The "voting" action is of course a metaphorical expression, and technically "1" is assigned to the corresponding address in the memory space allocated to each voxel in the voxel space.
Corresponding to storing. Such voting processing is performed on all the silhouette images B1 to B36.
As a result, the number of votes of voxels existing in a portion where all the hypothetical existing regions corresponding to 36 silhouette images overlap is “36”. In FIG. 6C, the voxels for which the voting process is performed and the number of votes is “36” are shown in black. By extracting only the voxels with 36 votes, the three-dimensional existence region of the object (subject) can be determined. Note that when extracting the voxels in which the subject exists, the voxels with the number of votes of “36” are not extracted, but the number of votes is greater than or equal to a predetermined threshold value (for example, 32) in consideration of the error. It is also possible to extract only the voxels and use them as the existing regions. By such voting processing, even if some of the plurality of silhouette images are inaccurate, the three-dimensional existence region can be determined with high accuracy.

After performing the voting process, the extracted voxels are expanded into polygons to create three-dimensional shape data of the subject. In the polygon creation, first, the three-dimensional existence area extracted as a result of the voting process is cut along a plurality of planes including the center axis of the polygon space (36 planes forming an angle of 10 degrees with the center axis), The contour line of each cut surface is calculated. Then, each contour line of each cut surface is approximated to a polygon, and the vertex coordinates of the polygon are obtained. Finally, the adjacent vertices are connected by a straight line, and the vertices corresponding to the contour lines of the respective cut surfaces are connected by a straight line between the adjacent cut surfaces to generate a polygon. By adjusting the approximation accuracy in the polygon approximation, the number of polygons finally formed can be adjusted. Note that the voting process and the polygon process are known, and for example, Japanese Patent Laid-Open No. 10-1
See 24704.

As described above, the three-dimensional shape data of the subject represented by the polygon is created, and the shape storage unit 14
stored in d. When the measurement site is the subject's bust, the three-dimensional shape data of the bust is stored in the shape storage unit 14d. Of course, three-dimensional shape data of the whole body of the subject including the bust can be created and stored in the shape storage unit 14d. Then, the three-dimensional shape data of the measured region is read from the shape storage unit 14d and the size of the measured region is calculated based on the three-dimensional shape data (S107). For example, in the case of a bust, the perimeter of the bust three-dimensional shape data stored in the storage unit may be calculated.

As described above, in this embodiment, the subject can be automatically measured by simply wearing the specific belt 16 on the measurement site while wearing the clothes, so that the subject's psychological burden is light and the expert It is possible to measure with excellent reproducibility and accuracy compared to when using a tape measure.

<Second Embodiment> In the first embodiment described above, the subject wears the belt 16 on the measurement site to obtain image data. However, in addition to the belt 16, the subject wears other members. Automatic measurement is also possible by taking a picture.

FIG. 7 shows an example of wearing equipment other than the belt 16. In this example, the subject is not specular, that is, non-specular (or perfectly diffuse)
Wearing the orthosis 20 having excellent adhesion. By wearing such an orthosis 20, the psychological burden on the subject is greatly reduced compared to the case of being naked, and the contour of the human body can be accurately imaged by using the contact member. In addition, by using the non-specular reflection member, it is possible to obtain image data in which the brightness level changes according to the distance between the subject and the camera 12 (that is, when there is unevenness, according to the degree of unevenness). , Three-dimensional shape data can be obtained based on this image data. An example of the non-specular reflection material is cloth. FIG. 7 shows a state in which the subject wears the orthosis 20 made of such a non-specular reflection material on the bust.

The computer 1 according to the present embodiment will be described below.
The three-dimensional shape data creation processing in 4 will be described. In the present embodiment, two cameras 12 are provided separated by a predetermined distance (stereo camera), and a subject is photographed by the two cameras 12 to acquire an image. The stereo image data obtained by photographing is supplied to the computer 14. The computer 14 creates three-dimensional shape data of the subject based on the parallax (stereo difference) of the stereo images obtained from the two cameras 12. Although three-dimensional shape data measurement of a target object based on parallax is known, this will also be briefly described.

FIG. 8 shows a conceptual diagram of generation of three-dimensional shape data by the stereo method. The relative positions of the camera 12R and the camera 12L are measured in advance and stored in the computer 14. The computer 14 obtains the correspondence relationship between the images obtained by the camera 12R (image R) and the images obtained by the camera 12L (image L). The correspondence is
It can be obtained by block matching. For example, the image is divided into blocks of 8 × 8 pixels, the image R and the image L are compared on a block-by-block basis, and the position where the total sum of the absolute values of the differences is the minimum is set as the optimum corresponding portion. Block comparisons can also be made within a limited range using epipolar constraints. The epipolar constraint assumes a three-dimensional straight line connecting the pixel P and the camera start point of the image R, and the corresponding point in the image L of one pixel P on the image R is such that the straight line is projected on the other image L. It is a constraint that it exists only on a straight line (epipolar line). In both images, the point corresponding to one point of the object is calculated from the image R and the image L, and then the three-dimensional position of one point of the object is calculated based on the triangulation principle from the positional relationship between the two cameras 12R and 12L. Can be calculated. Then, by performing such processing on a plurality of points of the object, a point group in a three-dimensional space is obtained, and by connecting adjacent points, three-dimensional surface data is generated.

In addition to block matching, the correspondence can be determined by extracting edges in the image and calculating the correspondence between the edges (block matching can also be included in the edge extraction). However, the problem of edge extraction is that it is not possible to determine the correspondence relationship between portions other than the edges of the image.

Therefore, as in this embodiment, the subject is asked to wear the orthosis 20 made of a non-specular reflection material, and the equiluminance line is calculated from the image by photographing this, and the edge is defined using this equiluminance line. Correspondence is determined for other parts as well. The isoluminance line is a curve obtained by connecting points having equal luminance in the image, and the isoluminance line 100 is illustrated in FIG. 9. The isoluminance line has a dynamic range of luminance 1 ≦
For t ≦ tmax, it is obtained by extracting a region having a brightness of t or more from the image and removing noise on the boundary line. Then, a rough correspondence is obtained between regions of equal luminance or higher, and further corresponding points are obtained between equal luminance lines. The epipolar constraint described above is used to find the corresponding points (“Reconstruction of curved surface by stereoscopic viewing of isoluminance lines” IEICE Transactions DI
I Vol. J77-D-II No. 9 pp. 16
73-1679). In this way, by combining the non-specular reflection material and the isoluminance line, it is possible to obtain three-dimensional shape data by using the stereo method even in a portion where no edge exists. Of course, both the edge method and the isoluminance line method can be used to extract corresponding points to generate three-dimensional shape data.

It should be noted that when the three-dimensional shape data is created by the stereo method, it is possible to easily search for corresponding points in the image R and the image L by another method. For example, as shown in FIG. 10, the subject wears a brace 22 having a specific mark on the measurement site, or a brace 24 having a random pattern as shown in FIG. 11 on the measurement site. In the case of FIG. 10, since a specific mark exists in the image R and the image L, three-dimensional shape data may be created by the stereo method using these marks as corresponding points. In addition, in the case of FIG. 11, since the random pattern functions as a marker (there is no repeating pattern, it is easy to search for corresponding points), and thus three-dimensional shape data can be similarly created.

Although the embodiment of the present invention has been described above, the present invention is not limited to this, and various modifications can be made. For example, in the present embodiment, the rotary table 1
Although the subject rides on the object 0 and shoots in all directions with the camera 12, multiple cameras 12 may be arranged around the subject to obtain an omnidirectional image with one shot.
FIG. 12 shows an arrangement example of such a camera 12. N (N ≧ 2) cameras 12 are arranged around the subject, and synchronous imaging is performed to acquire a plurality of images. At this time, N
The images captured by the two cameras are arranged so that the portions corresponding to each other are included, that is, the capturing ranges partially overlap each other. The N cameras 12 have their optical systems calibrated in advance using a reference object for calibration. The subsequent image processing is the same as that of the above-described embodiment. When a silhouette image is created by the chroma key technique, a background plate or a screen may be used to surround the subject and the camera 12 in a dome shape. It is also possible to arrange the cameras 12 in a cylindrical shape or a cubic shape.
The silhouette image may be created using the background difference (only the background is photographed in advance, and the background image is subtracted from the image obtained by photographing the subject). In this case, the background is a static environment, and it is necessary to make the background the same when photographing the background and when photographing the subject. When arranging a plurality of cameras 12 around the subject, it is difficult to determine the subject's existence area by the voting process using voxels if the number of the arranged cameras 12 is small. It is preferable to create.

Further, in the present embodiment, the influence of clothes is eliminated as much as possible by wearing the belt 16, but it is possible to make the clothes adhere to the body by other methods instead of the belt 16. For example, clothes can be brought into close contact with each other using wind pressure, static electricity, atmospheric pressure difference, or the like.

Further, when it is difficult to bring the clothes into close contact with each other, it is possible to irradiate strong light from the back of the subject and obtain a silhouette image of the clothes.

[0040]

As described above, according to the present invention, a subject can automatically perform measurement while wearing clothes, reduce the psychological burden on the subject, and perform measurement with excellent reproducibility and accuracy. it can.

[Brief description of drawings]

FIG. 1 is a system configuration diagram of an embodiment.

FIG. 2 is a configuration block diagram of a computer.

FIG. 3 is an explanatory diagram of a belt.

FIG. 4 is an explanatory diagram of a human body model when wearing a belt.

FIG. 5 is a processing flowchart of the embodiment.

FIG. 6 is an explanatory diagram of a voting process and a polygon creation process.

FIG. 7 is an explanatory view of a worn state of another embodiment.

FIG. 8 is an explanatory diagram of creating three-dimensional shape data by the stereo method.

FIG. 9 is an explanatory diagram of isoluminance lines.

FIG. 10 is a diagram illustrating a worn state of another embodiment.

FIG. 11 is an explanatory view of a worn state in still another state.

FIG. 12 is an explanatory diagram of another camera arrangement.

[Explanation of symbols]

10 rotating tables, 12 cameras, 14 computers.

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Kota Fujimura             2-5-3 Keihan Hondori, Moriguchi City, Osaka Prefecture             Within Yo Denki Co., Ltd. (72) Inventor Tomoya Terauchi             2-5-3 Keihan Hondori, Moriguchi City, Osaka Prefecture             Within Yo Denki Co., Ltd. F term (reference) 2F065 AA04 AA53 BB05 BB27 CC16                       FF04 FF05 JJ05 JJ26 MM04                       PP13 QQ24 QQ31                 5B050 BA07 BA09 BA12 DA07 EA02                       EA06 EA07 EA28                 5B057 AA18 BA02 BA13 CA08 CA13                       CA16 CB13 CB16 CE12 DB08                       DC02 DC09 DC16

Claims (7)

[Claims] 1. A method of measuring a subject in a clothing state, comprising the steps of photographing a subject with a belt wrapped around the clothing on the measuring region, and measuring the measured region from a plurality of images obtained by photographing. An automatic measuring method comprising: a step of creating three-dimensional shape data; and a step of obtaining the size of the measured region based on the three-dimensional shape data. 2. The method according to claim 1, wherein in the step of creating the three-dimensional shape data, the contour data of the subject in a plurality of images obtained by photographing the subject from a plurality of directions is used to determine a predetermined value. The three-dimensional shape data is created by calculating the existence probability of the subject in the voxel space. 3. A method of measuring a subject in a dressed state, the method comprising: taking an image of a subject wearing a brace made of a non-specular reflection material at the measuring portion; and measuring the measured portion from a plurality of images obtained by the photographing. And the step of obtaining the size of the measurement site based on the three-dimensional shape data. 4. A method of measuring a subject in a clothing state, the method comprising: photographing a subject wearing a brace having a specific mark on a measurement site; and measuring a plurality of images obtained by the photographing of the measurement site. An automatic measuring method comprising: a step of creating three-dimensional shape data; and a step of obtaining the size of the measured region based on the three-dimensional shape data. 5. A method for measuring a subject in a dressed state, the step of photographing a subject wearing a brace having a random pattern on a measuring region, and a step of photographing the measuring region from a plurality of images obtained by photographing. An automatic measuring method comprising: a step of creating three-dimensional shape data; and a step of obtaining the size of the measured region based on the three-dimensional shape data. 6. The method according to claim 3, wherein in the step of creating the three-dimensional shape data, two three-dimensional shape data are obtained by photographing at two photographing positions separated from each other by a predetermined distance. An automatic measuring method, characterized in that the three-dimensional shape data is created based on image parallax. 7. The automatic measuring method according to claim 6, wherein the parallax is calculated by using points on an isoluminance line in the image.