JP6593830B1 - Information processing apparatus, information processing method, dimension data calculation apparatus, and product manufacturing apparatus - Google Patents

Abstract

To provide highly accurate dimension data of an object. An information processing apparatus includes a conversion unit 24C and an estimation unit 24D. The conversion unit 24C may be a reception unit that receives a silhouette image of an object. The estimation unit 24D uses the object engine 21A that associates the silhouette image of the sample object and the values of a predetermined number of shape parameters associated with the sample object, and determines the shape parameter of the object from the received silhouette image. Estimate the value. Then, the estimated value of the shape parameter of the object is associated with dimension data related to an arbitrary part of the object. [Selection] Figure 1

Description

  The present disclosure relates to an information processing apparatus, an information processing method, a dimension data calculation apparatus, and a product manufacturing apparatus.

  Conventionally, an apparatus for manufacturing a product based on the shape of an object has been studied. For example, in Patent Document 1 (Japanese Patent Laid-Open No. 2017-018158), a fingernail is photographed to obtain a nail image, and a nail shape, a nail position, a nail curvature, and the like are attached based on the nail image. A technique for acquiring nail information necessary for nail creation and creating artificial nail parts based on the nail information is disclosed.

  However, in the prior art, it has been difficult to calculate the size and shape of the object with high accuracy.

  An information processing apparatus according to a first aspect uses a reception unit that receives a silhouette image of an object, and an object engine that associates a silhouette image of a sample object with values of a predetermined number of shape parameters associated with the sample object And an estimation unit for estimating the value of the shape parameter of the object from the received silhouette image, and the estimated shape parameter value of the object is converted into dimensional data related to an arbitrary part of the object. Associated. With such a configuration, the dimension data calculated for the object can be provided with high accuracy.

  In the information processing apparatus according to the second aspect, in the information processing apparatus according to the first aspect, a predetermined number of shape parameters associated with the sample object are obtained by reducing the dimension of the three-dimensional data of the sample object.

  The information processing apparatus according to the third aspect is the same as the information processing apparatus according to the second aspect, in which dimension reduction is performed by principal component analysis, and inverse projection of projection according to principal component analysis is performed on the estimated number of shape parameter values. Thus, the three-dimensional data of the object is calculated, and the three-dimensional data is associated with the dimension data.

  In the information processing device according to the fourth aspect, in the information processing device according to the third aspect, a predetermined number of principal components after the second order excluding the first order principal components are selected as the shape parameters.

  An information processing apparatus according to a fifth aspect is the information processing apparatus according to the fourth aspect, wherein the object is a person and the main component of the first order is associated with the height of the person.

  An information processing apparatus according to a sixth aspect is the information processing apparatus according to any one of the first to fifth aspects, wherein the silhouette image of the sample target is a predetermined direction in a three-dimensional object composed of three-dimensional data of the sample target It is a projection image.

  An information processing apparatus according to a seventh aspect is the information processing apparatus according to any one of the first aspect to the sixth aspect, wherein the object engine includes a silhouette image of the sample object and a predetermined number of shape parameters associated with the sample object. It is generated by learning the relationship with the value.

An information processing apparatus according to an eighth aspect is the information processing apparatus according to any one of the first to seventh aspects,
From the value of a predetermined number of shape parameters estimated for the object, three-dimensional data of a plurality of vertices in the object is configured, and based on the configured three-dimensional data, between any two vertices in the object And a calculation unit for calculating dimension data.

  In the information processing apparatus of the ninth aspect, in the calculation unit of the information processing apparatus of the eighth aspect, the dimension data between two vertices is on a three-dimensional object composed of three-dimensional data of a plurality of vertices in the target object. Calculated along the curved surface.

  An information processing apparatus according to a tenth aspect is the information processing apparatus according to any one of the first aspect to the ninth aspect, wherein the silhouette image of the object is based on depth data obtained using the depth data measurement apparatus. It is generated by separating an image and an image other than the object.

  An information processing apparatus according to an eleventh aspect is the information processing apparatus according to the tenth aspect, wherein the depth data measurement device is a stereo camera.

  An information processing method according to a twelfth aspect includes a step of learning a relationship between a silhouette image of a sample object and values of a predetermined number of shape parameters associated with the sample object, generating an object engine, and a silhouette of the object A step of receiving an image, a step of estimating a shape parameter value of the target object from the received silhouette image by using the target object engine, and a portion related to a part of the target object based on the value of the shape parameter of the target object Calculating dimension data to be processed. With such a configuration, the dimension data calculated for the object can be provided with high accuracy.

  An information processing apparatus according to a thirteenth aspect uses a reception unit that receives attribute data of an object, and an object engine that associates attribute data of the sample object with a predetermined number of shape parameter values associated with the sample object And an estimation unit that estimates the value of the shape parameter of the target object from the received attribute data, and the estimated shape parameter value of the target object is dimensional data related to an arbitrary part of the target object. Associated. With such a configuration, the dimension data calculated for the object can be provided with high accuracy.

  In the information processing apparatus according to the fourteenth aspect, in the information processing apparatus according to the thirteenth aspect, a predetermined number of shape parameters associated with the sample object are obtained by reducing the dimension of the three-dimensional data of the sample object.

  The information processing device according to the fifteenth aspect is the same as the information processing device according to the fourteenth aspect, in which dimension reduction is performed by principal component analysis, and a predetermined number of principal components after the second order excluding the first order principal components are the shape parameters. Selected.

  An information processing apparatus according to a sixteenth aspect is the information processing apparatus according to the fifteenth aspect, wherein the object is a person, the first principal component is associated with the height of the person, and the attribute data includes the height data of the object It is.

  An information processing apparatus according to a seventeenth aspect is the information processing apparatus according to any one of the thirteenth to sixteenth aspects, wherein the target object engine includes attribute data of the sample target object and a predetermined number of shape parameters associated with the sample target object. It is generated by learning the relationship with the value.

  An information processing apparatus according to an eighteenth aspect is the information processing apparatus according to any one of the thirteenth to seventeenth aspects, wherein three vertices of a plurality of vertices in the object are calculated based on a predetermined number of shape parameter values estimated for the object. And a calculation unit configured to calculate dimension data between two arbitrary vertices in the target object based on the configured three-dimensional data.

  In an information processing device according to a nineteenth aspect, in the calculation unit of the information processing device according to the eighteenth aspect, dimension data between two vertices is on a three-dimensional object composed of three-dimensional data of a plurality of vertices in the object. Calculated along the curved surface.

  An information processing method according to a twentieth aspect includes a step of learning a relationship between attribute data of a sample object and values of a predetermined number of shape parameters associated with the sample object, and generating an object engine; A step of receiving data, an estimation step of estimating a value of the shape parameter of the target object from the received attribute data, and a step of calculating dimension data related to a part of the target object based on the value of the shape parameter of the target object And including.

  A size data calculation apparatus according to a twenty-first aspect includes an acquisition unit that acquires image data obtained by photographing an object and full length data of the object, an extraction unit that extracts shape data indicating the shape of the object from the image data, and shape data. A silhouette image using a conversion unit that converts the image into a silhouette image based on the total length data, and an object engine that associates the silhouette image of the sample object with a predetermined number of shape parameter values associated with the sample object An estimation unit that estimates the value of a predetermined number of shape parameters, and a calculation unit that calculates dimensional data of the object based on the estimated value of the predetermined number of shape parameters. With such a configuration, the dimension data calculated for the object can be provided with high accuracy.

  In the dimensional data calculation device according to the twenty-second aspect, in the dimensional data calculation device according to the twenty-first aspect, the predetermined number of shape parameters are obtained by reducing the dimension of the three-dimensional data of the sample object.

  The dimensional data calculation apparatus according to the twenty-third aspect is the same as the dimensional data calculation apparatus according to the twenty-second aspect, in which dimension reduction is performed by principal component analysis, and the calculation unit projects a predetermined number of shape parameter values according to principal component analysis. Three-dimensional data is calculated by performing inverse transformation based on the matrix, and dimension data is calculated from the three-dimensional data.

  A product manufacturing apparatus according to a twenty-fourth aspect manufactures a product related to the shape of an object using at least one dimension data calculated by using any one of the dimension data calculation apparatuses according to any of the twenty-first to twenty-third aspects. .

It is a mimetic diagram of size data calculation system 100 concerning a 1st embodiment. It is a flowchart which shows operation | movement of the learning apparatus 25 of FIG. It is a flowchart which shows operation | movement of the dimension data calculation apparatus 20 of FIG. It is a schematic explanatory drawing which shows the characteristic of a shape parameter. It is a schematic graph which shows the characteristic of a shape parameter. It is a schematic diagram which shows the concept of the product manufacturing system 1 which concerns on 1st Embodiment. It is a sequence diagram which shows operation | movement of the product manufacturing system 1 of FIG. It is a schematic diagram which shows an example of the screen displayed on the terminal device 10 of FIG. It is a schematic diagram which shows an example of the screen displayed on the terminal device 10 of FIG. It is a schematic diagram of the dimension data calculation system 100 which concerns on 2nd Embodiment. It is a flowchart which shows operation | movement of the learning apparatus 125 of FIG. It is a flowchart which shows operation | movement of the dimension data calculation apparatus 20 of FIG. It is a schematic diagram which shows the concept of the product manufacturing system 1S which concerns on 2nd Embodiment.

Hereinafter, a dimension data calculation system according to embodiments of an information processing apparatus, an information processing method, a product manufacturing apparatus, and a dimension data calculation apparatus according to the present invention will be described with reference to the accompanying drawings. In the following description of the embodiments, the information processing apparatus and the dimension data calculation apparatus are implemented as a part of the dimension data calculation system.

  In the accompanying drawings, the same or similar elements are denoted by the same or similar reference numerals, and redundant description of the same or similar elements may be omitted in the description of each embodiment. Further, the features shown in each embodiment can be applied to other embodiments as long as they do not contradict each other. Furthermore, the drawings are schematic and do not necessarily match actual dimensions and ratios. In some cases, the dimensional relationships and ratios may be different between the drawings.

  In the following description, when a plurality of elements are described together using a matrix, a vector, or the like, they are expressed in capital letters, and in the case of expressing individual elements of a matrix, they may be expressed in lowercase letters. For example, when a set of shape parameters is indicated, it is expressed as a matrix Λ, and when an element of the matrix Λ is expressed, it may be expressed as an element λ.

<First Embodiment>
(1-1) Configuration of Dimension Data Calculation System FIG. 1 is a schematic diagram showing a configuration of a dimensional data calculation system 100 according to this embodiment. The dimension data calculation system 100 includes a dimension data calculation device 20 and a learning device 25.

  Each of the dimension data calculation device 20 and the learning device 25 can be realized by an arbitrary computer. The dimension data calculation device 20 includes a storage unit 21, an input / output unit 22, a communication unit 23, and a processing unit 24. The learning device 25 includes a storage unit 26 and a processing unit 27. The dimension data calculation device 20 and the learning device 25 may be realized as hardware using an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or the like.

  Each of the storage units 21 and 26 stores various types of information, and is realized by an arbitrary storage device such as a memory and a hard disk. For example, the storage unit 21 stores various data, programs, information, and the like including the target object engine 21 </ b> A in order for the processing unit 24 to perform information processing related to dimension data calculation. The storage unit 26 stores training data used in the learning stage in order to generate the object engine 21A.

The input / output unit 22 is realized by a keyboard, a mouse, a touch panel, and the like, and inputs various information to the computer and outputs various information from the computer.
The communication unit 23 is realized by a network interface such as an arbitrary network card, and enables communication with communication devices on the network by wire or wireless.

  The processing units 24 and 27 are each realized by a processor such as a CPU (Central Processing Unit) and / or a GPU (Graphical Processing Unit) and a memory in order to execute various types of information processing. The processing unit 24 functions as an acquisition unit 24A, an extraction unit 24B, a conversion unit 24C, an estimation unit 24D, and a calculation unit 24E by reading a program stored in the storage unit 21 into a CPU, GPU, or the like of a computer. . Similarly, the processing unit 27 functions as a preprocessing unit 27A and a learning unit 27B by reading a program stored in the storage unit 26 into a CPU, GPU, or the like of a computer.

  In the processing unit 24 of the dimension data calculation device 20, the acquisition unit 24A acquires image data obtained by photographing the object, and attribute data such as full length data and weight data of the object. The acquiring unit 24A acquires, for example, a plurality of image data obtained by capturing an object from a plurality of different directions using an imaging device.

  The extraction unit 24B extracts shape data indicating the shape of the object from the image data. Specifically, the extraction unit 24B extracts a target area included in the image data by using a semantic segmentation algorithm (Mask R-CNN or the like) prepared for each type of target object. Extract shape data. Semantic segmentation algorithms can be constructed using training data in which the shape of the object is not specified.

  Note that when the algorithm of semantic segmentation is constructed using training data of an object whose shape is unspecified, the shape of the object may not necessarily be extracted with high accuracy. In such a case, the extraction unit 24B extracts the shape data of the target object from the target object region using a grab cut algorithm. This makes it possible to extract the shape of the object with high accuracy.

  In addition, the extraction unit 24B separates the object and the background image other than the object by correcting the image of the object specified by the grab cut algorithm based on the color image of the specific part of the object. Also good. Thereby, it becomes possible to generate the shape data of the object with higher accuracy.

  The conversion unit 24C silhouettes the shape data based on the full length data. That is, the shape data of the object is converted to generate a silhouette image of the object. Thereby, shape data is normalized. The conversion unit 24C also functions as a reception unit for inputting the generated silhouette image to the estimation unit 24D.

  The estimation unit 24D estimates a predetermined number of shape parameter values from the silhouette image. The object engine 21A is used for the estimation. The value of the shape parameter of the predetermined number of objects estimated by the estimation unit 24D is associated with dimension data related to an arbitrary part of the object.

  The calculation unit 24E calculates the dimension data of the object associated with the predetermined number of shape parameter values estimated by the estimation unit 24D. Specifically, the calculation unit 24E configures three-dimensional data of a plurality of vertices in the object from the value of the shape parameter of the object estimated by the estimation unit 24D, and further, based on the three-dimensional data, the target Dimension data between any two vertices in the object is calculated.

  In the processing unit 27 of the learning device 25, the preprocessing unit 27A performs various preprocessing for learning. In particular, the preprocessing unit 27A specifies a predetermined number of shape parameters through feature extraction of the three-dimensional data of the sample object by dimension reduction. In addition, a predetermined number (dimensions) of shape parameter values are obtained for each sample object. The value of the shape parameter of the sample object is stored in the storage unit 26 as training data.

  In addition, the preprocessing unit 27A virtually configures a three-dimensional object of the sample target in the three-dimensional space based on the three-dimensional data of the sample target, and then virtually provides the three-dimensional space in the three-dimensional space. A three-dimensional object is projected from a predetermined direction using an imaging device. The silhouette image of the sample object is stored in the storage unit 26 as training data.

The learning unit 27B learns to associate the silhouette image of the sample object with the relationship between a predetermined number of shape parameter values associated with the sample object. As a result of learning, the object engine 21A is generated. The generated object engine 21A can be held in the form of an electronic file. When the dimension data calculation device 20 calculates the dimension data of the object, the object engine 21A is stored in the storage unit 21 and referenced by the estimation unit 24D.

(1-2) Operation of Dimension Data Calculation System The operation of the dimensional data calculation system 100 of FIG. 1 will be described with reference to FIGS. FIG. 2 is a flowchart showing the operation (S10) of the learning device 25, and generates the object engine 21A based on the sample object data. FIG. 3 is a flowchart showing the operation of the dimension data calculating device 20, and calculates the dimension data of the object based on the image data of the object.

(1-2-1) Operation of Learning Device First, sample object data is prepared and stored in the storage unit 26 (S11). In one example, the data to be prepared is data for 400 sample objects, including 5,000 3D data for each sample object. The three-dimensional data includes the three-dimensional coordinate data of the vertices of the sample object. Further, the three-dimensional data includes mesh information such as vertex information of each mesh constituting the three-dimensional object and the normal direction of each vertex, and full length data, weight data, and time-lapse data (including age and the like). Or other attribute data may be included.

  The three-dimensional data of the sample object is associated with a vertex number. In the above-described example, three-dimensional data of 5,000 vertices is associated with the vertex numbers # 1 to # 5,000 for each sample object. Also, all or part of the vertex numbers are associated with information on the part of the object. For example, when the object is “person”, the vertex number # 20 is associated with the “head vertex”, and similarly, the vertex number # 313 is “left shoulder shoulder peak”, and the vertex number # 521 is “right shoulder shoulder”. And so on.

Subsequently, the preprocessing unit 27A performs feature conversion into shape parameters by dimension reduction (S12). Specifically, for each sample object, feature extraction is performed on the three-dimensional data of the sample object by dimension reduction. As a result, a predetermined number (dimension number) of shape parameters are obtained. In one example, the number of dimensions of the shape parameter is 30. Dimension reduction is principal component analysis, Random Projection
It is realized by such a method.

  The preprocessing unit 27A converts the three-dimensional data into a predetermined number of shape parameter values for each sample object using a projection matrix for principal component analysis. Thereby, it is possible to remove noise from the three-dimensional data of the sample object and to compress the three-dimensional data while maintaining related characteristic information.

  As in the previous example, there are 400 sample object data, each sample object contains 3D (coordinate) data of 5,000 vertices, and each 3D data is characterized by 30-dimensional shape parameters Assume to convert. Here, a vertex coordinate matrix of [400 rows, 15,000 columns (5,000 × 3)] representing data of 400 sample objects is assumed to be a matrix X. Further, a projection matrix of [15,000 rows, 30 columns] generated by principal component analysis is set as a matrix W. By multiplying the vertex coordinate matrix X by the projection matrix W from the right, a matrix Λ that is a shape parameter matrix of [400 rows, 30 columns] can be obtained.

That is, the shape parameter matrix Λ can be calculated from the following formula.

As a result of the matrix calculation using the projection matrix W, the 15,000-dimensional data of each of the 400 sample objects is feature-converted into the 30-dimensional principal component shape parameters (λ ₁ ,..., Λ ₃₀ ). Is done. In the principal component analysis, the average value of 400 values (λ _1,1 ,..., Λ _400,1 ) for λ _i is calculated to be zero.

As a result of S12, after obtaining the shape parameter matrix Λ, the preprocessing unit 27A expands the shape parameter data set included in the shape parameter matrix Λ using random numbers (S13). In the above example, 400 data sets (λ _{i, 1} ,..., Λ _{i, 30} (1 ≦ i ≦ 400)) are converted into 10,000 shape parameter extended data sets (λ _{j , 1} ,..., Λ _{j, 30} (1 ≦ j ≦ 10,000)). Data expansion is performed using random numbers having a normal distribution. The extended data set has a normal distribution in which the variance of each shape parameter value is 3σ.

By performing inverse transformation based on the projection matrix W on the extension data set, the three-dimensional data of the extension data set can be configured. In the above example, an extended shape parameter matrix of [10,000 rows, 30 columns] representing 10,000 extended data sets is further transformed into a matrix Λ ′ (λ _{j, k} ,..., Λ _{j, k} (1 ≦ j ≦ 10,000 and 1 ≦ k ≦ 30)). The vertex coordinate matrix X ′ representing the three-dimensional data of 10,000 sample objects is a transposed matrix W ^T of the projection matrix W that is [30 rows, 15,000 columns] with respect to the extended shape parameter matrix Λ ′. Is obtained by multiplying from the right.

That is, the vertex coordinate matrix X ′ can be calculated from the following formula, and 5,000 (15,000 / 3) three-dimensional data are obtained for each 10,000 sample objects. be able to.

  As a result of S13, after the vertex coordinate matrix X 'is obtained, the pre-processing unit 27A generates respective silhouette images based on the expanded three-dimensional data of the sample object (S14). In the above-described example, a three-dimensional object of a sample object is virtually constructed from 5,000 pieces of three-dimensional data in a three-dimensional space for every 10,000 sample objects. Similarly, a three-dimensional object is projected using a projection device that is virtually provided in the three-dimensional space and can project from any direction. In the above-described example, it is preferable to obtain two silhouette images in the front direction and the side direction by projection for every 10,000 sample objects. The acquired silhouette image is represented by black and white binarized data.

  Finally, the learning unit 27B associates the relationship between the value of the shape parameter associated with the sample object and the silhouette image of the sample object through learning (S15). Specifically, it is preferable to learn the relationship between the shape parameter data set obtained in S13 and the silhouette image obtained in S14 as training data by deep learning.

More specifically, the binarized data of each silhouette image is input to the deep learning network architecture for the sample objects expanded to 10,000 in the above example. In feature extraction in deep learning, the network architecture weighting factor is set so that the data output from the network architecture approaches the values of 30 shape parameters. Note that the deep learning here can use, for example, a convolutional neural network (CNN).

  In this way, in S14, the relationship between the value of the shape parameter associated with the sample object and the silhouette image of the sample object is learned by deep learning, and a deep learning network architecture is constructed. As a result, the object engine 21A, which is an estimation model for estimating the value of the shape parameter, is generated in response to the silhouette image of the object being input.

(1-2-2) Operation of Dimension Data Calculation Device The dimensional data calculation device 20 previously stores the electronic file of the object engine 21A generated by the learning device 25 and the projection information of the principal component analysis obtained by the learning device 25. It is stored in the storage unit 21 and used to calculate the dimension data of the object.

  First, the acquisition unit 24A acquires a plurality of pieces of image data obtained by photographing the entire object from different directions via an external terminal device or the like, together with the full length data indicating the total length of the object, through the input / output unit 122 ( S21). Next, the extraction unit 24B extracts shape data indicating the shape of each part of the object from each image data (S22). Subsequently, the conversion unit 24C executes a rescaling process for converting each shape data into a predetermined size based on the full length data (S23). Through the steps S21 to S23, a silhouette image of the object is generated, and the dimension data calculation device 20 receives the silhouette image of the object.

  Next, using the object engine 21A stored in advance in the storage unit 21, the estimation unit 24D estimates the value of the shape parameter of the object from the received silhouette image (S24). Then, the calculation unit 24E calculates dimension data related to the part of the target object based on the value of the shape parameter of the target object (S25).

Specifically, in the calculation of the dimension data in S25, first, three-dimensional data of the vertices of the object is constructed from the values of a predetermined number of shape parameters estimated by the object engine 21A for the object. Here, the inverse transformation of the projection (S10) related to the dimension reduction performed by the preprocessing unit 27A in the learning stage may be performed. More specifically, with respect to the values of the shape parameters of the estimated predetermined number (column vector), by multiplying the transposed matrix W ^T of the projection matrix W according to the principal component analysis from the right, it is possible to obtain three-dimensional data.

In the above-described example, the three-dimensional data X ″ of the object can be calculated from the following mathematical formula for the value Λ ″ of the object shape parameter.

In S25, the calculation unit 24E calculates the dimension data between any two vertices in the object using the three-dimensional data. Here, a three-dimensional object is virtually constructed from three-dimensional data, and dimension data between two vertices are calculated along a curved surface on the three-dimensional object. That is, the distance between two vertices can be calculated three-dimensionally along the three-dimensional shape of the three-dimensional object. In order to calculate the distance between two vertices three-dimensionally, first, two vertices on a three-dimensional mesh composed of a large number of vertices (5,000 three-dimensional data in the above example) are used. Search for the shortest path connecting the two, and identify the mesh through which the shortest path passes. Next, using the vertex coordinate data of the identified mesh, the distance is calculated for each mesh along the shortest path, and is added together. The total value is a three-dimensional distance between the two vertices. In addition, mesh data such as vertex information of each mesh constituting the three-dimensional object and the normal direction of each vertex can be used for the calculation of the three-dimensional distance.

  For example, it is assumed that the object is “person” and the “shoulder width” of the person is calculated. As advance preparation, it is defined in advance that “shoulder width = the distance between the apex indicating the left shoulder shoulder peak and the apex indicating the right shoulder shoulder peak”. Further, the vertex number indicating the left shoulder shoulder peak is, for example, # 313, and the vertex number indicating the right shoulder shoulder peak is, for example, # 521. These pieces of information are stored in the storage unit 21 in advance. When calculating the dimension data, the shortest path from the vertex numbers # 313 to # 521 is specified, and the mesh vertex coordinate data specified in relation to the shortest path is used to determine the distance for each mesh along the shortest path. Can be calculated and summed.

  Thus, the dimension data calculation apparatus 20 of this embodiment can estimate the value of a predetermined number of shape parameters from a silhouette image with high accuracy by using the object engine 21A. In addition, since the three-dimensional data of the object can be restored with high accuracy from the value of the shape parameter estimated with high accuracy, high accuracy can be achieved between any two vertices as well as a specific part. Can be calculated. In particular, since the calculated dimension data between two vertices is calculated along a three-dimensional shape based on a three-dimensional object composed of three-dimensional data, it is highly accurate.

(1-3) Features of Dimension Data Calculation System a) As described above, the dimensional data calculation system 100 according to the present embodiment includes the dimensional data calculation device 20 and the learning device 25. The information processing apparatus configured as a part of the dimension data calculation apparatus 20 includes a conversion unit (accepting unit) 24C and an estimation unit 24D. The conversion unit (accepting unit) 24C receives a silhouette image of the object. The estimation unit 24D uses the object engine 21A that associates the silhouette image of the sample object and the values of a predetermined number of shape parameters associated with the sample object, and calculates the shape parameter of the object from the received silhouette image. Estimate the value. Then, the estimated value of the shape parameter of the object is associated with dimension data related to an arbitrary part of the object.

  The dimension data calculation device 20 includes an acquisition unit 24A, an extraction unit 24B, a conversion unit 24C, an estimation unit 24D, and a calculation unit 24E. The acquiring unit 24A acquires image data obtained by photographing the object and full length data of the object. The extraction unit 24B extracts shape data indicating the shape of the object from the image data. The converter 24C converts the shape data into a silhouette image based on the full length data. The estimation unit 24D uses the target object engine 21A that associates the silhouette image of the sample target object with the predetermined number of shape parameter values associated with the sample target object, and then calculates the predetermined number of shape parameter values from the silhouette image. presume. The calculation unit 24E calculates the dimension data of the object based on the estimated number of values of the predetermined number of shape parameters.

  Therefore, the dimension data calculation device 20 can estimate the value of a predetermined number of shape parameters from the silhouette image with high accuracy by using the object engine 21A created in advance. In addition, by using the shape parameter value estimated with high accuracy, it is possible to efficiently and highly accurately calculate data related to an arbitrary part of the object. Thus, according to the dimension data calculation device 20, the dimension data calculated for the object can be provided efficiently and with high accuracy.

By using such a dimension data calculation device 20, for example, the dimension data of each part of a living organism as an object can be calculated with high accuracy. Moreover, the dimension data of each part of an arbitrary object such as a car or various luggage can be calculated with high accuracy. Furthermore, by incorporating the dimension data calculation device 20 into a product manufacturing apparatus that manufactures various products, it is possible to manufacture a product that conforms to the shape of the object.

  By using such a dimension data calculation device 20, for example, dimension data related to each part of a living organism as an object can be calculated with high accuracy. Moreover, the dimension data of each part of an arbitrary object such as a car or various luggage can be calculated with high accuracy. Further, by incorporating the dimension data calculation device 20 into a product manufacturing apparatus that manufactures various products, it is possible to manufacture a product that conforms to the shape of the object.

  b) In the dimension data calculation device 20, a predetermined number of shape parameters associated with the sample object are specified by reducing the dimension of the three-dimensional data of the sample object. In particular, dimension reduction is performed by principal component analysis. Thereby, noise can be effectively removed from the three-dimensional data of the sample object, and the three-dimensional data can be compressed.

  c) In the dimension data calculation apparatus 20, the three-dimensional data of the object is calculated by inverse transformation of the projection according to the principal component analysis with respect to the estimated predetermined number of shape parameter values, and the three-dimensional data is the dimension data. Associated with Thereby, it is possible to construct the three-dimensional data of the target object with high accuracy with respect to the input of the silhouette image of the target object.

  d) In the dimension data calculation apparatus 20, the silhouette image of the sample object is a projection image in a predetermined direction on a three-dimensional object composed of the three-dimensional data of the sample object. That is, a three-dimensional object is constructed using the three-dimensional data of the sample object, and a silhouette image is obtained by projecting it. Two silhouette images in the front direction and the side direction are preferably acquired by projection. Thereby, the silhouette image of the sample object can be generated with high accuracy.

  e) In the dimension data calculation device 20, the object engine 21A is generated by learning the relationship between the silhouette image of the sample object and the values of a predetermined number of shape parameters associated with the sample object. The learning can be performed by deep learning. By performing learning by deep learning, the silhouette image of the sample object and the value of the shape parameter of the sample object can be associated with high accuracy.

  f) The calculation unit 24E of the dimension data calculation device 20 constructs three-dimensional data of a plurality of vertices in the target object from the predetermined number of shape parameter values estimated for the target object. Then, based on the configured three-dimensional data, dimensional data between any two vertices in the object is calculated. That is, after the three-dimensional object is constructed using the three-dimensional data of the object, the dimension data is calculated. Thereby, since the dimension data between two vertices can be calculated from the shape of the three-dimensional object of the target object, the measurement target part is not limited to a specific part.

  In particular, in the calculation unit 24E of the dimension data calculation device 20, dimension data between two vertices is calculated along a curved surface on a three-dimensional object composed of three-dimensional data of a plurality of vertices in the object. . Thereby, the dimension data can be calculated with higher accuracy.

(1-4) Modifications a) In the above description, the acquiring unit 24A acquires a plurality of image data obtained by capturing an object from different directions. Here, a depth data measuring device capable of acquiring depth data together is applicable as an imaging device that can simultaneously photograph an object from a plurality of different directions. An example of the depth data measuring device is a stereo camera. In this specification, “stereo camera”
The term “imaging device” means an arbitrary type of imaging device that captures a subject simultaneously from a plurality of different directions and reproduces binocular parallax. On the other hand, it is not always necessary to have a plurality of image data, and it is possible to calculate the dimension data of each part even if there is only one image data of the object.

b) In the above description, the grab cut algorithm may be adopted to separate the object and the background image other than the object. In addition to this, for example, when the above-described stereo camera is used in the acquisition unit 24A, it is possible to separate an object and a background image other than the object using depth data acquired from the stereo camera. It is. Thereby, it becomes possible to generate the shape data of the object with higher accuracy. In addition to the stereo camera, depth data can be obtained by a LiDAR (Light Detection and Ranging) device, and the object and the background image other than the object can be separated. That is, by using an arbitrary machine (depth data measuring device) capable of measuring the depth data for the acquiring unit 24A, it is possible to generate the shape data of the object with high accuracy.

  c) In the above description, in S13, the preprocessing unit 27A performs the data expansion process for extending the shape parameter data set using random numbers. In the data expansion process, it is only necessary to determine the number of samples to be expanded according to the number of sample objects. If the number of samples is sufficiently prepared in advance, the expansion process of S13 may not be performed.

c) In the above description, the shape parameters (λ ₁ ,..., λ ₃₀ ) are obtained by principal component analysis in the preprocessing unit 27A of the learning device 25 for the sample object. Here, the shape parameter when the sample object is “human” will be further considered. As a result of the principal component analysis using the three-dimensional data of 400 sample objects and 5,000 vertices as in the above example, the shape parameter when the object is “person” is at least It was considered to have characteristics.

[Characteristic 1]
The first-order principal component λ ₁ was related to have a linear relationship with the height of the person. Specifically, as shown in FIG. 4, the larger the first order principal component λ ₁ , the smaller the height of the person.

Considering the characteristics 1, when the estimated value of the shape parameter of the person by the estimation unit 24D, with respect the main component lambda ₁ of the first rank, the height data acquired by the acquisition unit 24A without using the object engine 21A Use it. Specifically, the value of the first rank principal component λ ₁ may be configured to be calculated separately by using a linear regression model having the human height as an explanatory variable.

In this case, even when generating an object engine 21A in the learning unit 27B in the learning stage, it may exclude the main component lambda ₁ from the learning target. As described above, in the learning stage of the learning device 25, the network architecture is weighted. At this time, there is an error between the principal components after the second rank excluding the first principal components obtained by performing the principal component analysis on the input silhouette image and the values after the shape parameter λ _{2 as} training data. The network architecture weighting factor may be set to be minimized. Thereby, when the target is “person”, the estimation accuracy of the value of the shape parameter in the estimation unit 24D can be improved together with the use of the linear regression model.

In addition, when the object engine 21A is generated in the learning unit 27B in the learning stage, the weight of the network architecture is minimized so that an error from the value after the shape parameter λ ₁ including the first rank principal component is minimized. A coefficient may be set. Then, the value of the first-order principal component λ ₁ may be replaced with a separately calculated value by using a linear regression model having the human height as an explanatory variable. Thereby, when the target is “person”, the estimation accuracy of the value of the shape parameter in the estimation unit 24D can be improved together with the use of the linear regression model.

[Characteristic 2]
FIG. 5 is a schematic graph showing the recall of the shape parameter as the main component. In the graph, the horizontal axis indicates the principal components ranked by the contribution rate, and the vertical axis indicates the variance explanation rate of the eigenvalues. And the bar graph has shown the individual distribution explanatory rate for every order. The solid line graph indicates the accumulation of the distributed explanation ratios from the first rank. In FIG. 5, for the sake of simplicity, a graph relating to ten principal components up to the tenth rank is schematically shown. Note that the eigenvalue of the covariance matrix obtained in the principal component analysis represents the magnitude of the eigenvector (principal component), and the variance explanation rate of the eigenvalue may be considered as a recall with respect to the principal component.

  Referring to the graph of FIG. 5, the cumulative explanation ratio of the 10 principal components from the first rank to the tenth rank is about 0.95 (broken arrow). That is, a person skilled in the art understands that the recall of the ten principal components from the first rank to the tenth rank is about 95%. That is, in the above-described example, the shape parameter subjected to feature conversion by dimension reduction is 30 dimensions. However, the shape parameter is not limited to this, and about 95% can be covered even if the shape parameter is 10 dimensions. That is, in the above description, the number of shape parameters (the number of dimensions) is 30. However, if the characteristic 2 is considered, the number may be about 10.

  d) In the above description, in the pre-processing unit 27A of the learning device 25, two silhouette images in the front direction and the side direction are obtained by projection for every 10,000 sample objects. On the other hand, two silhouette images of the object are not necessarily required, and only one is required.

(1-5) Application to Product Manufacturing System Hereinafter, an example in which the above-described dimension data calculation device 20 is applied to the product manufacturing system 1 will be described.

(1-5-1) Configuration of Product Manufacturing System FIG. 6 is a schematic diagram showing the concept of the product manufacturing system 1 according to the present embodiment. The product manufacturing system 1 includes a dimension data calculation device 20 that can communicate with the terminal device 10 owned by the user 5 and a product manufacturing device 30, and is a system for manufacturing a desired product 6. In FIG. 6, as an example, the concept when the object 7 is a person and the product 6 is a chair is shown. However, in the product manufacturing system 1 according to the present embodiment, the object 7 and the product 6 are not limited to these.

  The terminal device 10 can be realized by a so-called smart device. Here, the terminal device 10 exhibits various functions by installing a user program in the smart device. Specifically, the terminal device 10 generates image data captured by the user 5. Here, the terminal device 10 may have a stereo camera function of simultaneously capturing a subject from a plurality of different directions and reproducing binocular parallax. In addition, image data is not limited to what was image | photographed with the terminal device 10, For example, you may utilize what was image | photographed using the stereo camera installed in the shop.

  In addition, the terminal device 10 receives input of attribute data indicating the attribute of the object 7. The “attribute” includes the total length, weight, elapsed time from generation (including age), and the like of the object 7. Further, the terminal device 10 has a communication function, and executes transmission / reception of various information between the dimension data calculation device 20 and the product manufacturing device 30 and the terminal device 10.

  The dimension data calculation device 20 can be realized by an arbitrary computer. Here, the storage unit 21 of the dimension data calculation device 20 stores information transmitted from the terminal device 10 in association with identification information for identifying the user 5 of the terminal device 10. The storage unit 21 also stores parameters and the like necessary for executing information processing for calculating dimension data.

  Further, as described above, the processing unit 24 of the dimension data calculation device 20 functions as the acquisition unit 24A, the extraction unit 24B, the conversion unit 24C, the estimation unit 24D, and the calculation unit 24E. Here, the acquisition unit 24 </ b> A acquires the image data captured by the user 5 with the stereo camera and the attribute data of the object 7. Further, the extraction unit 24B extracts shape data indicating the shape of the object 7 from the image data. For example, when “person” is preset as the type of object, an algorithm for semantic segmentation is constructed using training data for identifying a person. In addition, the extraction unit 24B may separate the target object and the background image other than the target object using depth data acquired from the stereo camera. The converter 24C converts the shape data into a silhouette image based on the full length data. The conversion unit 24C also functions as a reception unit for inputting the generated silhouette image to the estimation unit 24D.

  The estimation unit 24D uses the target object engine 21A that associates the silhouette image of the sample target object with the predetermined number of shape parameter values associated with the sample target object, and then calculates the predetermined number of shape parameter values from the silhouette image. presume. The calculation unit 24E calculates the dimension data of the object based on the estimated number of values of the predetermined number of shape parameters. Specifically, three-dimensional data of a plurality of vertices in the target object is constructed from the value of the shape parameter of the target object estimated by the estimation unit 24D, and any two of the target object based on the three-dimensional data is further configured. Dimension data between two vertices is calculated.

  The product manufacturing apparatus 30 is a manufacturing apparatus that manufactures a desired product related to the shape of the object 7 using at least one dimension data calculated using the dimension data calculating apparatus 20. The product manufacturing apparatus 30 can employ any apparatus that can automatically manufacture and process a product, and can be realized by, for example, a three-dimensional printer.

(1-5-2) Operation of Product Manufacturing System FIG. 7 is a sequence diagram for explaining the operation of the product manufacturing system 1 according to the present embodiment. 8 and 9 are schematic diagrams showing screen transitions of the terminal device 10.

  First, a plurality of images are captured so that the entire object 7 is captured from different directions via the terminal device 10, and a plurality of image data in which the object 7 is captured is generated (T1). Here, a plurality of front and side photographs as shown in FIGS. 8 and 9 are taken. Such front and side photographs should be taken with the stereo camera function of the terminal device 10 turned on.

  Next, attribute data indicating the attribute of the object 7 is input to the terminal device 10 by the user 5 (T2). Here, full length data, weight data, time-dependent data (including age, etc.) of the object 7 are input as attribute data. The plurality of image data and attribute data are transmitted from the terminal device 10 to the dimension data calculation device 20.

  When the dimension data calculation device 20 receives a plurality of image data and attribute data from the terminal device 10, the dimension data calculation device 20 calculates the dimension data of each part of the object 7 using these data (T3). The terminal device 10 displays dimension data on the screen according to the setting. Then, the product manufacturing apparatus 30 manufactures a desired product 6 based on the dimension data calculated by the dimension data calculation apparatus 20 (T4).

(1-5-3) Features of Product Manufacturing System As described above, the product manufacturing system 1 according to this embodiment includes the dimension data calculation device 20 that can communicate with the terminal device 10 owned by the user 5, and the product manufacturing. Device 30.

  The terminal device 10 (imaging device) captures a plurality of images of the object 7. The dimension data calculation device 20 includes an acquisition unit 24A, an extraction unit 24B, a conversion unit 24C, an estimation unit 24D, and a calculation unit 24E. The acquiring unit 24A acquires image data obtained by photographing the object and full length data of the object. The extraction unit 24B extracts shape data indicating the shape of the object from the image data. The converter 24C converts the shape data into a silhouette image based on the full length data. The estimation unit 24D uses the target object engine 21A that associates the silhouette image of the sample target object with the predetermined number of shape parameter values associated with the sample target object, and then calculates the predetermined number of shape parameter values from the silhouette image. presume. The calculation unit 24E calculates the dimension data of the object based on the estimated number of values of the predetermined number of shape parameters. The product manufacturing apparatus 30 manufactures the product 6 using the dimension data calculated by the calculation unit 24E. With such a configuration, the dimension data calculation device 20 calculates each part of the object 7 with high accuracy, so that a desired product related to the shape of the object 7 can be provided.

  For example, an organ model can be manufactured by measuring the shape of various organs such as the heart by the product manufacturing system 1. In addition, for example, various health care products can be manufactured from measurement of a person's waist shape. For example, the figure product of the person can be manufactured from the shape of the person. In addition, for example, a chair or the like suitable for the person can be manufactured from the shape of the person. Also, for example, a car toy can be manufactured from the shape of the car. Further, for example, a diorama or the like can be manufactured from an arbitrary landscape image.

  In the above description, the dimension data calculation device 20 and the product manufacturing device 30 are described as separate devices, but they may be configured integrally.

Second Embodiment
Hereinafter, substantially the same reference numerals are given to the configurations and functions already described, and the description thereof is omitted.
(2-1) Configuration of Dimension Data Calculation Device FIG. 10 is a schematic diagram showing a configuration of a dimensional data calculation system 200 according to this embodiment. The dimension data calculation system 200 includes a dimension data calculation device 120 and a learning device 125.

  The dimension data calculation device 120 includes a storage unit 121, an input / output unit 122, a communication unit 123, and a processing unit 124. The learning device 125 includes a storage unit 126 and a processing unit 127. The dimension data calculation device 120 and the learning device 125 may be realized as hardware using an LSI, an ASIC, an FPGA, or the like.

  Each of the storage units 121 and 126 stores various types of information, and is realized by an arbitrary storage device such as a memory and a hard disk. For example, the storage unit 121 stores various data, programs, information, and the like including the target object engine 121 </ b> A in order for the processing unit 124 to perform information processing related to dimensional data calculation. In addition, the storage unit 126 stores training data used in the learning stage in order to generate the object engine 121A.

  The input / output unit 122 has the same configuration and function as the input / output unit 22 described above. The communication unit 123 has the same configuration and function as the communication unit 23 described above.

The processing unit 124 reads the program stored in the storage unit 121 into a CPU, GPU, or the like of a computer, thereby obtaining an acquisition unit 124A, an estimation unit 124D, and a calculation unit 124.
Functions as E. Similarly, the processing unit 127 functions as a preprocessing unit 127A and a learning unit 127B by reading a program stored in the storage unit 126 into a CPU, GPU, or the like of a computer.

  In the processing unit 124 of the dimension data calculation device 120, the acquisition unit 124A acquires attribute data including at least one of the full length data, the weight data, and the temporal data (including age and the like) of the target object. In the present embodiment, the acquisition unit 124A also functions as a reception unit for inputting attribute data to the estimation unit 24D.

  The estimation unit 124D estimates a predetermined number of shape parameter values from the attribute data. The object engine 121A is used for the estimation. The value of the shape parameter of the object estimated by the estimation unit 124D can be associated with dimension data related to an arbitrary part of the object, as will be described later.

  The calculation unit 124E calculates the dimension data of the object from the value of the shape parameter of the object estimated by the estimation unit 124D. Specifically, the calculation unit 124E configures three-dimensional data of a plurality of vertices in the target object from the value of the shape parameter of the target object estimated by the estimation unit 124D, and further, based on the three-dimensional data Dimension data between any two vertices in the object is calculated.

  In the processing unit 127 of the learning device 125, the preprocessing unit 127A performs various preprocessing for learning. In particular, the preprocessing unit 127A specifies a predetermined number of shape parameters through feature extraction of the three-dimensional data of the sample object by dimension reduction. The value of the shape parameter of the sample object and the corresponding attribute data are stored in advance in the storage unit 126 as training data.

  It is assumed that corresponding attribute data (full length data, weight data, time-dependent data (including age, etc.)) is prepared as the three-dimensional data of the sample object. Corresponding attribute data is stored in the storage unit 126 as training data.

  The learning unit 127B learns to associate the relationship between the value of the shape parameter of the sample object and the corresponding attribute data. As a result of learning, an object engine 121A is generated. The generated object engine 121A can be held in the form of an electronic file. When the dimension data calculation device 120 calculates the dimension data of the object, the object engine 121A is stored in the storage unit 121 and referenced by the estimation unit 124D.

(2-2) Operation of Dimension Data Calculation System The operation of the dimensional data calculation system 200 of FIG. 10 will be described with reference to FIGS. FIG. 11 is a flowchart showing the operation (S110) of the learning device 125. Here, the object engine 121A is generated based on the sample object data. FIG. 12 is a flowchart showing the operation of the dimension data calculation device 120. Here, the dimension data of the object is calculated based on the image data of the object.

(2-2-1) Operation of Learning Device First, sample object data is prepared and stored in the storage unit 126 (S111). In one example, the prepared data is data of 400 sample objects, and 5,000 pieces of three-dimensional data prepared for each sample object and attribute data prepared for each sample object are included. Including. The three-dimensional data includes the three-dimensional coordinate data of the vertices of the sample object. Further, the three-dimensional data may include mesh data such as vertex information of each mesh constituting the three-dimensional object and a normal direction of each vertex.

  Similarly to the first embodiment, the three-dimensional data of the sample object is associated with the part information together with the vertex number.

  Subsequently, the preprocessing unit 127A performs feature conversion into a predetermined number (dimensions) of shape parameters by dimension reduction (S112). This feature conversion process is also the same as in the first embodiment. In the above example, as a result of matrix calculation using the projection matrix for principal component analysis, the 15,000-dimensional (5,000 × 3) data of each of the 400 sample objects is, for example, a 30-dimensional principal component Is transformed into a shape parameter Λ.

  Then, the learning unit 127B uses the combination of the attribute data of the plurality of sample objects prepared in S111 and the data set of the plurality of shape parameters obtained in S112 as training data, and machine learning the relationship between the two (S115).

Specifically, the learning unit 27B obtains conversion attribute data Y from the attribute data of the target object. The element of the transformation matrix Z that associates the element y _r of the transformation attribute data Y with the element λ _m of the shape parameter Λ is denoted as z _rm . The transformation matrix Z is a matrix composed of [s rows and n columns]. The symbol m is 1 ≦ m ≦ n. In the above example, n is 30, which is the number of dimensions of the shape parameter Λ. The symbol r is 1 ≦ r ≦ s, and s is the number of elements used in the calculation obtained from the conversion attribute data Y.

  For example, it is assumed that the attribute data of the object includes full length data h, weight data w, and time-lapse data a. That is, the attribute data is a set of elements (h, w, a). In this case, the learning unit 27B multiplies each element (h, w, a) of the attribute data of the target object (h, w, a) to a square (also referred to as a secondary term) and a value obtained by multiplying each element (also referred to as an interaction term). And the value of each element itself (also referred to as a primary term).

As a result, conversion attribute data Y having the following nine elements is obtained.

  Next, the learning unit 27B performs regression analysis on a set of transformation attribute data Y obtained from attribute data associated with 400 sample objects and shape parameters Λ obtained from three-dimensional data of the sample objects. Thus, a transformation matrix Z composed of [9 rows, 30 columns] as shown below is obtained.

このようにして求められた変換行列Ｚのデータは、対象物エンジン１２１Ａとして記憶部２６に記憶される。

The data of the transformation matrix Z obtained in this way is stored in the storage unit 26 as the object engine 121A.

(2-2-2) Operation of Dimension Data Calculation Device The dimension data calculation device 120 stores the electronic file of the object engine 121A generated by the learning device 125 and the projection information of the principal component analysis obtained by the learning device 125. It is stored in the unit 121 and used to calculate the dimension data of the object.

  First, the acquiring unit 124A acquires the attribute data of the object through the input / output unit 122 (S121). Thereby, the attribute data of the object is received. Next, using the object engine 121A stored in advance in the storage unit 121, the estimation unit 124D estimates the value of the shape parameter of the object from the received attribute data (S124).

  For example, it is assumed that the attribute data of the object includes full length data h, weight data w, and time-lapse data a. That is, the attribute data is a set of elements (h, w, a). In S124, as described above, the estimation unit 124D calculates a value obtained by squaring each element (h, w, a) of the attribute data of the target object, a value obtained by multiplying each element, and a value of each element itself. Conversion attribute data Y consisting of

  Finally, the calculation unit 124E calculates dimensional data related to the part of the object based on the value of the shape parameter of the object (S125). Specifically, the conversion attribute data Y is calculated from the attribute data of the object acquired by the acquisition unit 124A. Then, the shape parameter Λ is calculated by multiplying the conversion attribute data Y by the conversion matrix Z described above. Thereafter, as in the first embodiment (S25), a three-dimensional object is virtually constructed from three-dimensional data, and dimension data between two vertices are calculated along a curved surface on the three-dimensional object. To do. In addition, mesh data such as vertex information of each mesh constituting the three-dimensional object and the normal direction of each vertex can be used for the calculation of the three-dimensional distance.

  Thus, the dimension data calculation apparatus 120 of this embodiment can estimate the value of the predetermined number of shape parameters from the attribute data of the object with high accuracy by using the object engine 121A. Unlike the first embodiment, there is no need to input an image of an object, and S22 (shape data extraction processing) and S23 (rescaling processing) in FIG. 3 are not required, which is efficient.

  In addition, since the three-dimensional data of the object can be restored with high accuracy from the value of the shape parameter estimated with high accuracy, high accuracy can be achieved between any two vertices as well as a specific part. Can be calculated. In particular, since the calculated dimension data between two vertices is calculated along a three-dimensional shape based on a three-dimensional object composed of three-dimensional data, it is highly accurate.

(2-3) Features of Dimension Data Calculation System As described above, the dimension data calculation system 200 according to the present embodiment includes the dimension data calculation device 120 and the learning device 125. The information processing apparatus configured as a part of the dimension data calculation apparatus 120 includes an acquisition unit (accepting unit) 124A, an estimation unit 124D, and a calculation unit 124E. The acquisition unit (accepting unit) 124A receives the attribute data of the object. The estimation unit 124D uses the object engine 121A in which the attribute data of the sample object and the values of a predetermined number of shape parameters associated with the sample object are associated, and the object shape parameter is determined from the received attribute data. Estimate the value of. Then, the estimated value of the shape parameter of the object is associated with dimension data related to an arbitrary part of the object.

Therefore, the dimension data calculation apparatus 120 can efficiently estimate the value of a predetermined number of shape parameters from the attribute data by using the object engine 121A that has been created in advance. The estimated shape parameter value is highly accurate. In addition, by using the shape parameter value estimated with high accuracy, it is possible to efficiently and highly accurately calculate data related to an arbitrary part of the object. Thus, according to the dimension data calculation apparatus 120, the dimension data calculated about an object can be provided efficiently and with high accuracy.

(2-4) Application to Product Manufacturing System FIG. 13 is a schematic diagram showing a concept of a product manufacturing system 1S according to the present embodiment. The dimension data calculation device 120 according to the present embodiment can also be applied to the product manufacturing system 1S, similarly to the dimension data calculation device 20 according to the first embodiment.

  The terminal device 10 </ b> S according to the present embodiment only needs to accept input of attribute data indicating the attribute of the object 7. Examples of the “attribute” include the total length, weight, and elapsed time (including age) from the generation of the object 7.

  Further, as described above, the processing unit 124 of the dimension data calculation device 120 functions as the acquisition unit 124A, the estimation unit 124D, and the calculation unit 124E. The calculation unit 124E calculates dimension data related to a part of the target object based on the value of the shape parameter of the target object obtained by the estimation unit 124D.

  In the product manufacturing system 1S, since the dimension data calculation device 120 calculates the dimension data of the object 7 efficiently and with high accuracy, a desired product related to the shape of the object 7 can be provided. In addition, the product manufacturing system 1S according to the second embodiment can exhibit the same effects as the product manufacturing system 1 of the first embodiment.

<Other embodiments>
The present disclosure is not limited to the above embodiments as they are. The present disclosure can be embodied by modifying the components without departing from the scope of the disclosure in the implementation stage. Further, the present disclosure can form various disclosures by appropriately combining a plurality of constituent elements disclosed in the respective embodiments. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements may be appropriately combined in different embodiments.

DESCRIPTION OF SYMBOLS 1,1S ... Product manufacturing system, 100, 200 ... Dimension data calculation system, 20, 120 ... Dimension data calculation apparatus, 21, 26, 121, 126 ... Storage part, 21A, 121A ... Object engine, 22, 122 ... input / output unit, 23,123 ... communication unit, 24,27,124,127 ... processing unit, 24A, 124A ... acquisition unit, 24B ... extraction unit, 24C ... conversion unit, 24D, 124D ... estimation unit, 24E, 124E ... Calculation unit, 25, 125 ... Learning device, 27A, 127A ... Pre-processing unit, 27B, 127B ... Learning unit

JP 2017-018158 A

Claims (11)

A receiving unit for receiving attribute data including at least one of the total length data and the weight data of the object;
And attribute data including at least one of the full length data and the weight data of the sample object, using an object engine which associates the value of the shape parameters of a predetermined number associated with the sample object, An estimation unit for estimating a value of a shape parameter of the object from the received attribute data;
An information processing apparatus, wherein the estimated shape parameter value of the target object is associated with dimensional data related to an arbitrary part of the target object.   The attribute data corresponding to the three-dimensional data of the sample object is associated with the predetermined number of shape parameter values specified by characteristic conversion of the three-dimensional data in the object engine. The information processing apparatus described. The information processing apparatus according to claim 1 or 2 , wherein a predetermined number of shape parameters associated with the sample object are obtained by reducing the dimensions of the three-dimensional data of the sample object. The information processing apparatus according to claim 3 , wherein the dimension reduction is performed by principal component analysis, and a predetermined number of principal components after the second order excluding the first order principal components are selected as the shape parameter. The information processing apparatus according to claim 4, wherein the object is a person, the principal component of the first rank is associated with a person's height, and full length data of the attribute data corresponds to the person 's height data.   The information processing apparatus according to any one of claims 1 to 5, wherein the object and the sample object are people. The object engine, the sample object attribute data that is generated by learning the relationship between the value of the shape parameter of the predetermined number associated with the sample object, any one of claims 1 to 6 The information processing apparatus according to one item. From the estimated shape parameter value for the object, three-dimensional data of a plurality of vertices in the object is constructed, and any two of the vertices in the object are constructed based on the constructed three-dimensional data The information processing apparatus according to any one of claims 1 to 7 , further comprising a calculation unit that calculates dimensional data between the two. In the calculating unit, dimension data between the two vertices, the calculated along the curved surface of the three-dimensional object composed of three-dimensional data of the plurality of vertices in the object, according to claim 8 Information processing device. And attribute data including at least one of the full length data and the weight data of the sample object, and the values of the shape parameters of the predetermined number associated with the sample object, a relationship is learned, the Target product engine Generating step;
Receiving attribute data including at least one of full length data and weight data of the object;
An estimation step of estimating a value of a shape parameter of the object from the received attribute data;
Calculating dimension data relating to a part of the object based on a value of a shape parameter of the object;
Including an information processing method.   The object engine is configured to associate the attribute data corresponding to the three-dimensional data of the sample object with the values of the predetermined number of shape parameters specified by characteristic conversion of the three-dimensional data. 10. The method according to 10.

Images