JP2008171074A - Three-dimensional shape model generation device, three-dimensional shape model generation method, computer program, and three-dimensional model generation system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a three-dimensional shape model generation device, a three-dimensional shape model generation method, a computer program and a three-dimensional shape model generation system capable of surely generating a fine three-dimensional shape model on which geometrical features unique to a modeling object are reflected even when the modeling object is not included in a three-dimensional shape model information group. <P>SOLUTION: The three-dimensional shape model information of one group of a plurality of modeling objects is generated based on the three-dimensional shape data of those modeling objects. A plurality of featured values are extracted by carrying out multivariate analysis of the generated three-dimensional shape model information. The plurality of extracted featured values are stored, and the three-dimensional shape model information belonging to the predetermined one group is generated based on the values of the featured values. When accepting the input of the three-dimensional shape data of the modeling object excluding the object of multivariate analysis, the three-dimensional shape model information is generated based on the three-dimensional shape data to predict the plurality of featured values corresponding to the three-dimensional shape model information. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a three-dimensional shape model generation apparatus and a three-dimensional shape model generation method for accurately generating a three-dimensional shape model based on a small number of measurement values acquired by a simple measurement method without using a method such as three-dimensional digitization. , A computer program, and a three-dimensional shape model generation system. In particular, even for a modeled object that has a complicated shape such as a human body and whose shape features between individuals vary widely, a simple measurement method such as dimension measurement is used. The present invention relates to a three-dimensional shape model generation apparatus, a three-dimensional shape model generation method, a computer program, and a three-dimensional shape model generation system capable of generating a highly accurate three-dimensional shape model.

  With the rapid development of computer technology in recent years, various applications using a digitized three-dimensional shape model have been developed. And there are objects that have unique features in the modeled objects, and how to precisely generate a three-dimensional shape model that expresses individual features is an important issue. .

  As a means for generating a precise three-dimensional shape model, a method based on measurement data by an optical measurement device using a measurement device such as a laser is generally used. These measurement data can reproduce a three-dimensional shape model with high accuracy.

  However, the number of data measured by a normal measuring device is several hundred thousand to several million, and the number of data, topology, etc. are different among the measurement data. In addition, the format of measurement data differs depending on the type of measurement device, and there is a problem that post-processing such as statistical analysis becomes complicated. In addition, optical measuring devices using lasers are expensive and complicated to handle, so they cannot be used easily in stores, homes, etc., and obstruct the spread of applications using 3D shape models. There was also a problem that it was a factor.

  In order to solve such a problem, for example, an arbitrary three-dimensional shape model generation method by making a homologous model as shown in Patent Document 1 is known. Patent Document 1 discloses an example in which an object including many personal features in a shape like a human body is used as a modeled object. In other words, a 3D shape model of a human body with a standard body shape is generated as a standard model, the measurement data measured by the measurement device is aligned with the standard model, the scale is matched, the posture is aligned, etc. By executing the processing, homologous three-dimensional shape model data (hereinafter referred to as three-dimensional shape model information) in which individual shape characteristics are expressed is generated.

By generating 3D shape model information, the number of data, topology, etc. are the same with much fewer data points (eg, several thousand points) than the measurement data without losing the shape characteristics of the 3D shape based on the measurement data. It is possible to generate precise three-dimensional shape model information consisting of a unified data format. Therefore, the three-dimensional shape model information can be easily statistically analyzed. For example, by performing multivariate analysis such as principal component analysis (PCA analysis), information (modeling) of the entire three-dimensional shape model information group is performed. When the object is a human body, information regarding whether or not the object is large, whether it is a solid body shape, or the like) can be extracted (see Patent Document 1 and Non-Patent Document 1).
Japanese Patent No. 3364654 B. Allen and two others, “Exploring the Space of Human Body Shapes: Data-driven Synthesis Under Anthropometeric Control”, SAE Design and Design Digital Human Modeling Symposium 2004 (SAE Digital Human Modeling for Design and Engineering Symposium 2004), p.2004-01-2188, 2004 Akinori Nagata, Toru Okazaki, Masaaki Tsuji, Hiroshi Harashima, “Regular generation and spatial description of facial images by principal component analysis”, IEICE Transactions, Vol.J79-D-II No.7, p.1230-1235, 1996

  For example, when the object to be modeled is a human body, when a population composed of a plurality of 3D shape model information is used as a 3D shape model information group representing a Japanese body shape, the population is subjected to multivariate analysis. For example, by performing principal component analysis or the like, it is possible to extract body shape characteristic components of Japanese. The extracted feature components are referred to as feature amounts or principal components.

  When a predetermined human body is selected from the population, the feature value (principal component value) of the selected human body can be calculated. The position in the feature amount space (principal component space) can be estimated. Therefore, conversely, when the feature value (principal component value) is specified, the three-dimensional shape model information can be restored (Non-Patent Document 2).

  In addition, for human bodies included in the population, by performing multivariate analysis, for example, principal component analysis, etc., body features can be packaged as eigenvector matrices, and each person's body shape Expressed as a value (principal component value). In general, features with low contributions can be omitted, and each person's body shape can be expressed with a data amount much smaller than the original data amount, and the original body shape can be reproduced with high accuracy. .

  However, when acquiring the shape data or feature value of a human body not included in the population, the following problems occur. First, the shape data needs to be measured using, for example, a laser type three-dimensional shape measurement scanner. However, these measuring devices are generally expensive, and when the object to be modeled is a human body, there is a problem that the measuring load on the subject is large, such as the subject needs to be naked. In addition, since the human body that is the modeled object is a moving object, it is difficult to remain stationary during measurement, and the measurement data for the number of points necessary to construct a three-dimensional shape model must be obtained accurately. There was a problem that it was difficult.

  On the other hand, feature values cannot be obtained directly unless they are included in the population, and must be estimated indirectly based on values measured by hand, measurement data obtained by a simple measurement method, etc. I must. However, if the modeled object to be estimated is greatly deviated from the population, for example, if the modeled object is a foreigner when the population is Japanese, the value of the feature amount of the human body The (principal component value) cannot be estimated accurately. Further, even in the case of the same Japanese, the distribution of feature amounts differs due to differences in age, gender, occupation, birthplace, and the like. Therefore, in order to accurately estimate the feature value, it is possible to improve the estimation accuracy by preparing multivariate analysis results of various categories.

  The present invention has been made in view of such circumstances, and even if it is a modeled object that is not included in the prepared three-dimensional shape model information group, it reflects geometric features unique to the modeled object. It is an object of the present invention to provide a three-dimensional shape model generation device, a three-dimensional shape model generation method, a computer program, and a three-dimensional shape model generation system that can reliably generate a detailed three-dimensional shape model.

  Further, even when a simple measuring means that does not cause a measurement load on the subject is used, it is possible to generate a precise three-dimensional shape model that reflects a shape characteristic unique to a modeling target such as a human body. It is an object to provide a three-dimensional shape model generation apparatus, a three-dimensional shape model generation method, a computer program, and a three-dimensional shape model generation system.

  In order to achieve the above object, a three-dimensional shape model generation device according to the first invention generates a group of three-dimensional shape model information of a modeled object based on three-dimensional shape data of a plurality of modeled objects. Three-dimensional shape model information generating means for performing, multi-variate analysis of the three-dimensional shape model information generated by the three-dimensional shape model information generating means to extract a plurality of feature amounts, and the feature amount extraction A plurality of feature value values extracted by the means and three-dimensional shape model information stored in association with each other, and three-dimensional shape model information belonging to a predetermined group is generated based on the feature value. In a 3D shape model generation device comprising a 3D shape model generation means and an output means for outputting the generated 3D shape model information, the 3D shape of the modeling object that is not subject to multivariate analysis 3D shape data receiving means for receiving data input, and a 3D shape model for generating 3D shape model information of the modeled object based on the 3D shape data received by the 3D shape data receiving means An information generation means and a means for estimating a plurality of feature value values corresponding to the three-dimensional shape model information with reference to the analysis result storage means.

  Further, in the first invention, the three-dimensional shape model generation apparatus according to the second invention is a group of three-dimensional shape information according to the condition for a plurality of three-dimensional shape model information that matches a predetermined condition for the same type of modeling object. It is generated as shape model information, and includes condition receiving means for receiving information on the predetermined condition, and the feature amount extracting means is a group corresponding to the information on the predetermined condition received by the condition receiving means. The three-dimensional shape model information is selected, and a plurality of feature amounts are extracted by multivariate analysis of the selected group of three-dimensional shape model information.

  In addition, the 3D shape model generation apparatus according to the third aspect of the invention is a 3D shape model information generation unit that generates 3D shape model information of a modeled object based on 3D shape data of a plurality of modeled objects. Means, feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the three-dimensional shape model information generated by the three-dimensional shape model information generation means, and a plurality of features extracted by the feature quantity extraction means Analysis result storage means for storing the feature value and the three-dimensional shape model information in association with each other, three-dimensional shape model generation means for generating the three-dimensional shape model information based on the feature value, and the generated tertiary In a three-dimensional shape model generation apparatus comprising an output means for outputting original shape model information, a physical quantity receiving means for receiving input of information relating to a predetermined physical quantity capable of measuring a modeling target The three-dimensional shape includes correspondence relation calculating means for calculating a correspondence relation between the information related to the predetermined physical quantity received by the physical quantity receiving means and the values of the plurality of feature quantities extracted by the feature quantity extraction means. The model generation unit is configured to obtain the values of the plurality of feature amounts extracted by the feature amount extraction unit based on the information related to the predetermined physical amount received by the physical amount reception unit and the correspondence relationship calculated by the correspondence relationship calculation unit. And a means for inquiring of the analysis result storage means based on a plurality of feature value values estimated by the means to generate new three-dimensional shape model information.

  According to a fourth aspect of the present invention, there is provided a three-dimensional shape model generating apparatus that generates three-dimensional shape model information of a modeled object based on three-dimensional shape data of a plurality of modeled objects. Means, feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the three-dimensional shape model information generated by the three-dimensional shape model information generation means, and a plurality of features extracted by the feature quantity extraction means Analysis result storage means for storing the feature value and the three-dimensional shape model information in association with each other, three-dimensional shape model generation means for generating the three-dimensional shape model information based on the feature value, and the generated tertiary In a three-dimensional shape model generation apparatus comprising an output means for outputting original shape model information, a physical quantity receiving means for receiving input of information relating to a predetermined physical quantity capable of measuring a modeling target , Virtual 3D that stores virtual 3D shape model information generated by changing each feature quantity step by step in association with a predetermined physical quantity that can be measured and a plurality of feature quantity values Shape model information storage means, and the three-dimensional shape model generation means inquires of the virtual three-dimensional shape model information storage means based on information about the predetermined physical quantity received by the physical quantity reception means, and A means for specifying a plurality of feature quantity values extracted by the feature quantity extraction means, and interpolating a plurality of feature quantity values specified by the means to estimate a feature quantity value corresponding to the modeled object And means for inquiring of the analysis result storage means based on the feature value estimated by the means and generating new three-dimensional shape model information.

  Further, the 3D shape model generation method according to the fifth aspect of the invention is generated by generating 3D shape model information of a group of the modeling objects based on the 3D shape data of a plurality of modeling objects. Multivariate analysis of the 3D shape model information to extract a plurality of feature amounts, store the extracted plurality of feature amount values and the 3D shape model information in association with each other, and determine based on the feature value 3D shape model information that belongs to a group and outputs the generated 3D shape model information. In the 3D shape model generation method that outputs the generated 3D shape model information, Receiving an input, generating three-dimensional shape model information of the modeled object based on the received three-dimensional shape data, and estimating a plurality of feature values corresponding to the three-dimensional shape model information, To do.

  The three-dimensional shape model generation method according to the sixth invention is the method according to the fifth invention, wherein a plurality of three-dimensional shape model information that matches a predetermined condition is obtained from a group of three-dimensional shapes according to conditions for the same type of modeling object. Generated as shape model information, receives information on the predetermined condition, selects a corresponding group of 3D shape model information based on the received information on the predetermined condition, and selects the selected group of 3D shape model information. It is characterized by extracting a plurality of feature values by multivariate analysis.

  Further, the 3D shape model generation method according to the seventh invention generates 3D shape model information of the modeled object based on the 3D shape data of a plurality of modeled objects, and generates the generated 3D Multivariate analysis of the shape model information to extract a plurality of feature values, store the extracted feature value values in association with the 3D shape model information, and 3D shapes based on the feature value values In the three-dimensional shape model generation method for generating model information and outputting the generated three-dimensional shape model information, an input of information on a predetermined physical quantity that can be measured of the modeling target is received, and the received predetermined The correspondence between the information on the physical quantity and the extracted values of the plurality of feature quantities is calculated, and the extracted information on the plurality of feature quantities is calculated based on the received information on the predetermined physical quantity and the calculated correspondence relation. To estimate, based on the value of the estimated plurality of feature amounts, and generates a new three-dimensional shape model information.

  Further, the 3D shape model generation method according to the eighth invention generates 3D shape model information of the modeled object based on the 3D shape data of the plurality of modeled objects, and generates the generated 3D Multivariate analysis of the shape model information to extract a plurality of feature values, store the extracted feature value values in association with the 3D shape model information, and 3D shapes based on the feature value values In the three-dimensional shape model generation method that generates model information and outputs the generated three-dimensional shape model information, it accepts input of information related to a predetermined physical quantity that can be measured of the modeling object, and each feature quantity The virtual three-dimensional shape model information generated by stepwise changing is stored in association with a predetermined physical quantity that can be measured and a plurality of feature quantity values, and the received predetermined physical quantity is related to Based on the information, the value of the extracted plurality of feature values is specified, the value of the specified feature values is interpolated, and the value of the feature value corresponding to the modeled object is estimated and estimated. Based on the feature value, new three-dimensional shape model information is generated.

  A computer program according to a ninth aspect of the invention is a computer program that generates three-dimensional shape model information for a group of modeled objects based on three-dimensional shape data of a plurality of modeled objects. Generating means, feature quantity extracting means for extracting a plurality of feature amounts by multivariate analysis of the three-dimensional shape model information generated by the three-dimensional shape model information generating means, and a plurality of features extracted by the feature quantity extracting means Analysis result storage means for associating and storing quantity values and three-dimensional shape model information, three-dimensional shape model generation means for generating three-dimensional shape model information belonging to a predetermined group based on feature value values, and generation In a computer program that can be executed by a computer that functions as output means for outputting the generated three-dimensional shape model information, the computer includes: 3D shape data receiving means for receiving input of 3D shape data of a modeling object that is not subject to variable analysis, and the modeling object based on the 3D shape data received by the 3D shape data receiving means The three-dimensional shape model information generating means for generating the three-dimensional shape model information and the analysis result storage means are referred to function as means for estimating a plurality of feature value values corresponding to the three-dimensional shape model information. It is characterized by.

  A computer program according to a tenth aspect of the invention is the computer program according to the ninth aspect, wherein a plurality of three-dimensional shape model information that matches a predetermined condition for a same type of modeling object is obtained from a group of three-dimensional shape information according to the condition. A function generating unit configured to function as a shape model information and a condition receiving unit configured to receive information related to the predetermined condition, and the feature amount extracting unit is configured to correspond to a group of corresponding information based on the information related to the predetermined condition received by the condition receiving unit. It is characterized by selecting three-dimensional shape model information and functioning as a means for extracting a plurality of feature amounts by multivariate analysis of the selected group of three-dimensional shape model information.

  A computer program according to an eleventh aspect of the invention is a computer program for generating a 3D shape model information generating means for generating 3D shape model information of a modeled object based on 3D shape data of a plurality of modeled objects. , Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the three-dimensional shape model information generated by the three-dimensional shape model information generation means; a plurality of feature quantities extracted by the feature quantity extraction means; Analysis result storage means for associating and storing values and 3D shape model information, 3D shape model generation means for generating 3D shape model information based on feature value values, and generated 3D shape model information In a computer program that can be executed by a computer that functions as an output means for outputting the computer, the computer measures the modeled object. Physical quantity accepting unit that accepts input of information related to a predetermined physical quantity, correspondence relationship between information about the predetermined physical quantity accepted by the physical quantity accepting unit, and values of a plurality of feature quantities extracted by the feature quantity extracting unit A plurality of feature quantities extracted by the feature quantity extraction means based on the correspondence relation calculation means for calculating the information, information on the predetermined physical quantity received by the physical quantity acceptance means, and the correspondence relation calculated by the correspondence relation calculation means And a function for inquiring the analysis result storage means based on a plurality of feature value values estimated by the means and generating new three-dimensional shape model information. And

A computer program according to a twelfth aspect of the invention is a computer program for generating 3D shape model information generating means for generating 3D shape model information of a modeled object based on 3D shape data of a plurality of modeled objects. , Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the three-dimensional shape model information generated by the three-dimensional shape model information generation means; a plurality of feature quantities extracted by the feature quantity extraction means; Analysis result storage means for associating and storing values and 3D shape model information, 3D shape model generation means for generating 3D shape model information based on feature value values, and generated 3D shape model information In a computer program that can be executed by a computer that functions as an output means for outputting the computer, the computer measures the modeled object. A physical quantity accepting unit that accepts input of information related to a predetermined physical quantity, a predetermined physical quantity capable of measuring virtual three-dimensional shape model information generated by changing each feature quantity in stages, And the virtual three-dimensional shape model information storage means for storing the information in association with the values of the plurality of feature amounts, and inquiring the virtual three-dimensional shape model information storage means based on the information on the predetermined physical quantity received by the physical quantity reception means. Means for specifying a plurality of feature quantity values extracted by the feature quantity extraction means, and interpolating a plurality of feature quantity values specified by the means to obtain a feature quantity value corresponding to the modeled object And a means for querying the analysis result storage means based on the feature value estimated by the means and generating new three-dimensional shape model information. To.

  A three-dimensional shape model generation system according to a thirteenth aspect of the present invention is the three-dimensional shape model generation system including a central device and a terminal device connected so as to be capable of data communication via a network. Three-dimensional shape model information generating means for generating a group of three-dimensional shape model information based on three-dimensional shape data of a plurality of modeled objects, and the three-dimensional shape model information generating means A feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the three-dimensional shape model information generated in step (a), a plurality of feature quantity values extracted by the feature quantity extraction means, and a three-dimensional shape model information; Analysis result storage means for storing information relating to the correspondence relationship, three-dimensional shape data receiving means for receiving the three-dimensional shape data of the modeled object that is not subject to multivariate analysis, Based on the three-dimensional shape data received by the three-dimensional shape data receiving means, referring to the three-dimensional shape model information generating means for generating the three-dimensional shape model information of the modeled object, and the analysis result storage means, Means for estimating a plurality of feature value values corresponding to the three-dimensional shape model information, a plurality of feature value values estimated by said means, and information relating to correspondence stored in said analysis result storage means; A feature amount transmitting means for transmitting to the apparatus, wherein the terminal device acquires a three-dimensional shape data acquisition means for acquiring three-dimensional shape data of a modeled object that is not subject to multivariate analysis; and the acquired three-dimensional shape Three-dimensional shape data transmitting means for transmitting data to the central apparatus, receiving means for receiving information on the estimated values of the plurality of feature values and correspondences from the central apparatus, and the received plural A unit for generating three-dimensional shape model information based on the feature value and information on the correspondence, and an output unit for outputting a plurality of feature value and reconstructed three-dimensional shape model information. And

  The three-dimensional shape model generation system according to a fourteenth aspect of the present invention is the system according to the thirteenth aspect, wherein the central device sets a plurality of pieces of three-dimensional shape model information matching a predetermined condition for the same type of modeling object Separately, it is generated as a group of three-dimensional shape model information, and the terminal device receives a condition receiving unit that receives information related to the predetermined condition, and information related to the predetermined condition received by the condition receiving unit And the central device comprises means for receiving information on the predetermined condition from the terminal device, and the feature quantity extracting means is based on the information on the predetermined condition received by the means. Select a corresponding group of 3D shape model information and extract multiple features by multivariate analysis of the selected group of 3D shape model information It characterized that you have.

  A three-dimensional shape model generation system according to a fifteenth aspect of the present invention is the three-dimensional shape model generation system including a central device and a terminal device connected so as to be able to perform data communication via a network. Is generated by the three-dimensional shape model information generating means for generating the three-dimensional shape model information of the modeled object based on the three-dimensional shape data of the plurality of modeled objects, and the three-dimensional shape model information generating means. Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the obtained 3D shape model information, and correspondence between a plurality of feature quantity values extracted by the feature quantity extraction means and the 3D shape model information Analysis result storage means for storing information relating to the relationship, physical quantity receiving means for receiving information relating to a predetermined physical quantity that can be measured of the modeled object, and reception of the physical quantity Correspondence calculation means for calculating the correspondence between the information about the predetermined physical quantity received in the stage and the values of the plurality of feature quantities extracted by the feature quantity extraction means, and the predetermined physical quantity received by the physical quantity reception means Based on the information and the correspondence calculated by the correspondence calculation means, means for estimating the values of the plurality of feature quantities extracted by the feature quantity extraction means, and a plurality of feature quantities estimated by the means Characteristic amount transmitting means for transmitting information about values and correspondence stored in the analysis result storage means to the terminal device, wherein the terminal device is a predetermined physical quantity capable of measuring the modeled object Information on physical quantity acquisition means for acquiring information on the physical quantity, physical quantity transmission means for transmitting information on the acquired physical quantity to the central device, and information on the estimated values of the plurality of feature quantities and the correspondence relationship Receiving means from the central device, means for generating three-dimensional shape model information based on the received information about the values of the plurality of feature quantities and the correspondence relationship, the values of the plurality of feature quantities and the reconstructed tertiary Output means for outputting original shape model information.

  The three-dimensional shape model generation system according to the sixteenth aspect of the present invention is the three-dimensional shape model generation system comprising a central device and a terminal device connected so as to be capable of data communication via a network. Is generated by the three-dimensional shape model information generating means for generating the three-dimensional shape model information of the modeled object based on the three-dimensional shape data of the plurality of modeled objects, and the three-dimensional shape model information generating means. Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the obtained 3D shape model information, and correspondence between a plurality of feature quantity values extracted by the feature quantity extraction means and the 3D shape model information Analysis result storage means for storing information relating to the relationship, physical quantity receiving means for receiving information relating to a predetermined physical quantity that can be measured of the modeled object, Virtual three-dimensional shape model information storage that stores virtual three-dimensional shape model information generated by changing the amount step by step in association with predetermined physical quantities that can be measured and values of a plurality of feature amounts And the virtual three-dimensional shape model information storage means based on the information related to the predetermined physical quantity received by the physical quantity receiving means, and specifies a plurality of feature quantity values extracted by the feature quantity extracting means Means for interpolating a plurality of feature value values identified by the means to estimate a feature value value corresponding to the modeled object, and a plurality of feature value values estimated by the means And feature amount transmission means for transmitting information related to the correspondence relationship stored in the analysis result storage means to the terminal device, wherein the terminal device relates to a predetermined physical quantity that can be measured by the modeled object. information Obtained physical quantity acquisition means, physical quantity transmission means for transmitting information on the acquired physical quantity to the central apparatus, receiving means for receiving information on the estimated values of the plurality of feature quantities and correspondences from the central apparatus, and received Means for generating three-dimensional shape model information based on information about a plurality of feature values and correspondence, and output means for outputting a plurality of feature values and reconstructed three-dimensional shape model information It is characterized by.

  In the first invention, the fifth invention, and the ninth invention, based on the three-dimensional shape data of the plurality of modeling objects, a group of three-dimensional shape model information is generated, and the generated tertiary A plurality of feature quantities are extracted by multivariate analysis of the original shape model information. A plurality of extracted feature value values and 3D shape model information are stored in association with each other, and 3D shape model information belonging to a predetermined group is generated based on the feature value values. When generating three-dimensional shape model information, it accepts input of three-dimensional shape data of a modeling target that is not subject to multivariate analysis, and based on the received three-dimensional shape data, the three-dimensional shape of the modeling target Shape model information is generated, and a plurality of feature value values corresponding to the three-dimensional shape model information are estimated. As a result, for example, even when 3D shape data not included in the population of 3D shape model information that has been homologized in advance is received, the value of the feature amount can be easily estimated. Based on the estimated feature value, it is possible to clearly identify where the modeling object that was not included in the population is positioned in the multidimensional space at the time of multivariate analysis.

  In the thirteenth invention, the central device generates a group of three-dimensional shape model information of the modeled object based on the three-dimensional shape data of the plurality of modeled objects, and the generated three-dimensional shape model information Are subjected to multivariate analysis to extract a plurality of feature amounts, and information regarding the correspondence relationship between the extracted values of the plurality of feature amounts and the three-dimensional shape model information is stored. The terminal device acquires the three-dimensional shape data of the modeled object that is not subject to multivariate analysis, and transmits it to the central device. The central device receives the 3D shape data, generates 3D shape model information of the modeled object, and estimates a plurality of feature value values corresponding to the 3D shape model information. Information about the estimated values of the plurality of feature amounts and the stored correspondence relationship is transmitted to the terminal device. The terminal device receives information on the estimated values of the plurality of feature amounts and the correspondence relationship, and generates and outputs 3D shape model information. As a result, for example, even when 3D shape data not included in the population of 3D shape model information that has been homologized in advance is received, the value of the feature amount can be easily estimated. Based on the estimated feature value, it is possible to clearly identify where the modeling target that was not included in the population is positioned in the multidimensional space after the multivariate analysis. Further, even when a computer having a high processing capacity is not employed as the terminal device, the three-dimensional shape model information can be easily generated, and the cost of the entire system can be reduced.

  It should be noted that the distribution between the processing in the central device and the processing in the terminal device can change so as to maintain the balance of the entire system. For example, as a result of multivariate analysis in the terminal device, the correspondence between the feature value and the 3D shape model information is stored, and only the feature value specified from the central device is received and the 3D shape model is received. By generating information, communication traffic can be reduced.

  In the second invention, the sixth invention, the tenth invention, and the fourteenth invention, a plurality of three-dimensional shape model information that matches a predetermined condition for the same modeled object as a group of three-dimensional shape model information according to the conditions. Remember. A corresponding group of three-dimensional shape model information is selected based on information on conditions, and a plurality of feature amounts are extracted by performing multivariate analysis on the selected group of three-dimensional shape model information. As a result, for example, feature quantities can be extracted based on a group of 3D shape model information corresponding to conditions such as Japanese, foreigners, men, women, adults, children, etc. Shape model information can be generated.

  In the third invention, the seventh invention, and the eleventh invention, three-dimensional shape model information of the modeled object is generated based on the three-dimensional shape data of the plurality of modeled objects, and the generated three-dimensional shape Multivariate analysis of model information to extract multiple features. A plurality of extracted feature value values and 3D shape model information are stored in association with each other, and 3D shape model information is generated based on the feature value values. When generating 3D shape model information, it accepts information about a predetermined physical quantity that can be measured of the modeling target, and correspondence between the received information about the predetermined physical quantity and the extracted values of multiple feature quantities Calculate the relationship. Based on the received information on the predetermined physical quantity and the calculated correspondence, the value of the extracted plurality of feature quantities is estimated, and based on the estimated value of the plurality of feature quantities, a new three-dimensional shape model Generate information. Thereby, based on physical quantities that can be measured by some means even when it is not possible to obtain detailed three-dimensional shape data and, for example, a homologous model cannot be configured as the three-dimensional shape model information. Thus, by obtaining the correspondence relationship with the feature value, the feature value can be easily estimated, and precise three-dimensional shape model information can be generated. The correspondence between physical quantities and feature values that can be measured by any means may be calculated as a conversion function or may be calculated as a conversion table.

  In the fifteenth aspect, the central device generates three-dimensional shape model information of the modeled object based on the three-dimensional shape data of the plurality of modeled objects, and performs multivariate analysis to obtain a plurality of feature values. Information regarding the correspondence relationship between the extracted feature value values and the three-dimensional shape model information is stored. The terminal device acquires information on a predetermined physical quantity that can be measured of the modeled object, and transmits the information to the central device. The central device receives information related to the predetermined physical quantity, and calculates a correspondence relationship between the received information related to the predetermined physical quantity and the extracted values of the plurality of feature quantities. The central device estimates values of the plurality of extracted feature values based on the calculated correspondence relationship, and transmits information about the estimated feature value values and the correspondence relationship to the terminal device. The terminal device receives information on the estimated values of the plurality of feature amounts and the correspondence relationship from the central device, and generates and outputs three-dimensional shape model information. Thereby, based on physical quantities that can be measured by some means even when it is not possible to obtain detailed three-dimensional shape data and, for example, a homologous model cannot be configured as the three-dimensional shape model information. Thus, by obtaining the correspondence relationship with the feature value, the feature value can be easily estimated, and precise three-dimensional shape model information can be generated. The correspondence between physical quantities and feature values that can be measured by any means may be calculated as a conversion function or may be calculated as a conversion table.

  It should be noted that the distribution between the processing in the central device and the processing in the terminal device can change so as to maintain the balance of the entire system. For example, as a result of multivariate analysis in the terminal device, the correspondence between the feature value and the 3D shape model information is stored, and only the feature value specified from the central device is received and the 3D shape model is received. By reconfiguring information, it is possible to reduce communication traffic.

  In the fourth invention, the eighth invention, and the twelfth invention, three-dimensional shape model information of the modeled object is generated based on the three-dimensional shape data of the plurality of modeled objects, and the generated three-dimensional shape The model information is subjected to multivariate analysis to extract a plurality of feature amounts, and the extracted plurality of feature amount values and the three-dimensional shape model information are stored in association with each other. When generating three-dimensional shape model information, virtual three-dimensional shape model information generated by receiving information on a predetermined physical quantity that can be measured for the modeling target and changing each feature value stepwise Are stored in association with predetermined physical quantities that can be measured and a plurality of feature quantity values. Based on the received information on the predetermined physical quantity, the value of the extracted plurality of feature quantities is specified, and the value of the plurality of specified feature quantities is interpolated, and the value of the feature quantity corresponding to the modeled object Is estimated. New three-dimensional shape model information is generated based on the estimated feature value. As a result, even if it is not possible to obtain detailed three-dimensional shape data and a homologous model cannot be configured as three-dimensional shape model information, for example, physical quantities and features that can be measured by some means By obtaining a correspondence table with quantity values in advance, it is possible to easily estimate the value of the feature quantity and to generate precise three-dimensional shape model information. The feature value may be estimated by, for example, using a lookup table and interpolating from the feature value values of a plurality of inquired three-dimensional shape model information.

  In the sixteenth invention, the central device generates three-dimensional shape model information of the modeled object based on the three-dimensional shape data of the plurality of modeled objects, and performs multivariate analysis to obtain a plurality of feature values. The information about the correspondence between the extracted feature value and the three-dimensional shape model information is stored. The terminal device acquires information on a predetermined physical quantity that can be measured of the modeled object, and transmits the information to the central device. The central device receives information related to a predetermined physical quantity, and a predetermined physical quantity capable of measuring virtual three-dimensional shape model information generated by changing each feature quantity stepwise, and a plurality of features It is stored in association with the amount value. The central device identifies a plurality of extracted feature value values based on the received information on the predetermined physical quantity, interpolates the identified plurality of feature value values, and features corresponding to the modeled object Estimate the value of the quantity. Information about the estimated values of the plurality of feature amounts and the stored correspondence relationship is transmitted to the terminal device. The terminal device receives information on the estimated values of the plurality of feature amounts and the correspondence relationship, and generates and outputs three-dimensional shape model information. As a result, even if it is not possible to obtain detailed three-dimensional shape data and a homologous model cannot be configured as three-dimensional shape model information, for example, physical quantities and features that can be measured by some means By obtaining a correspondence table with quantity values in advance, it is possible to easily estimate the value of the feature quantity and to generate precise three-dimensional shape model information. The feature value may be estimated by, for example, using a lookup table and interpolating from the feature value values of a plurality of inquired three-dimensional shape model information.

  It should be noted that the distribution between the processing in the central device and the processing in the terminal device can change so as to maintain the balance of the entire system. For example, as a result of principal component analysis, the terminal device stores the correspondence between the feature value and the three-dimensional shape model information, and receives only the feature value specified from the central device to obtain the three-dimensional shape model information. By generating, communication traffic can be reduced.

  According to the present invention, for example, even when 3D shape data not included in the population of 3D shape model information that has been homologized in advance is received, the value of the feature amount is easily estimated. Based on the estimated feature value, it is possible to clearly identify where the modeling target that was not included in the population is positioned in the multidimensional space at the time of multivariate analysis. it can. In addition, even if it is not possible to obtain detailed 3D shape data and a homologous model cannot be configured as 3D shape model information, for example, physical quantities and feature quantities that can be measured by some means By previously obtaining a correspondence table with the value of the feature value, it is possible to easily estimate the value of the feature value, and to generate precise three-dimensional shape model information. Therefore, it is possible to estimate 3D shape data more easily, and it is possible to contribute to the spread of various applications utilizing the 3D shape model in stores, homes, etc., such as electronic commerce of clothing, etc. Has an effect.

(Embodiment 1)
Hereinafter, the three-dimensional shape model generation apparatus according to Embodiment 1 of the present invention will be specifically described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a three-dimensional shape model generation apparatus 1 according to Embodiment 1 of the present invention. In FIG. 1, a three-dimensional shape model generation apparatus 1 includes at least a CPU (central processing unit) 11, an auxiliary storage unit 12, a RAM 13, a storage unit 14, a communication unit 15, an input unit 16, a display unit 17, an output unit 18, And an internal bus 19 for connecting the hardware described above.

  The CPU 11 is connected to the above-described hardware units of the three-dimensional shape model generation apparatus 1 via the internal bus 19, and controls the above-described hardware units and stores the computer program 3 stored in the RAM 13. Various software functions are executed according to the above. The RAM 13 includes an SRAM, a flash memory, and the like, and stores a computer program (load module), temporary data generated when the computer program is executed, and the like.

  The storage unit 14 includes a built-in fixed storage device (hard disk), a ROM, and the like, and a portable recording medium 2 such as a DVD or a CD-ROM. The computer program 3 stored in the storage means 14 is downloaded by the auxiliary storage means 12 from the portable recording medium 2 such as a CD-ROM in which information such as programs and data is recorded, and from the storage means 14 to the RAM 13 at the time of execution. Deploy and execute. In addition, the storage unit 14 includes an analysis result storage unit 141 that stores an analysis result obtained by performing multivariate analysis on the three-dimensional shape model information of the modeling target to be modeled.

  When the modeled object is a human body, for example, three-dimensional shape model information obtained by homogenizing a plurality of three-dimensional shape data representing a Japanese body shape is generated and subjected to multivariate analysis. Note that the homogenized three-dimensional shape model information means three-dimensional shape data having the same number of landmark points and the same geometric structure and having feature points that are anatomically associated with each other. Yes. Accordingly, the number of data points, topology, etc. are the same among the three-dimensional shape data, and the data format is unified.

  FIG. 2 is an exemplary diagram showing shape feature points of the human body when the modeled object is a human body. As shown in FIG. 2, the shape feature points of the human body are the trochanter points P1 and P2, the shoulder peak points P3 and P4, the nipple points P5 and P6, and the like, among the three-dimensional shape data points of the human body. It means a point with characteristics.

  When the object to be modeled is a human body, the analysis result storage unit 141 performs a multivariate analysis on a plurality of three-dimensional shape model information representing, for example, a Japanese body shape, and extracts a plurality of feature amounts. When the principal component analysis is performed, it is stored as an eigenvector matrix, average value, standard deviation, etc. constituting a plurality of extracted feature quantities. According to the stored eigenvector matrix, average value, standard deviation, etc., it is possible to generate 3D shape model information corresponding to an arbitrary position in the multidimensional feature amount space.

  The communication means 15 is connected to an internal bus 19 and can transmit / receive data to / from an external computer by being connected to an external network such as the Internet, LAN, or WAN. Therefore, it is possible to use the three-dimensional shape model information stored in one or a plurality of external computers.

  The input unit 16 is a device that acquires three-dimensional shape data of a modeled object, and is a laser-type or optical non-contact three-dimensional scanner. Thereby, hundreds of thousands to several millions of three-dimensional shape data necessary for generating a precise three-dimensional shape model can be acquired.

  The display unit 17 is a display device such as a CRT monitor or LCD, and displays the 3D shape data acquired by the input unit 16 or the generated 3D shape model. The output means 18 is a printing device such as a laser printer or a dot printer.

  Hereinafter, the operation of the three-dimensional shape model generation apparatus 1 configured as described above will be described. The three-dimensional shape model generation apparatus 1 first generates three-dimensional shape model information based on a plurality of standard three-dimensional shape data. For example, when the object to be modeled is a human body, a plurality of three-dimensional shape data of a standard body shape of the human body is homogenized, and a plurality of three-dimensional shape model information is generated as, for example, a polygon model. The standard three-dimensional shape data is composed of three-dimensional shape data measured with high accuracy by, for example, a well-known laser type three-dimensional shape measuring device, scanning device, or the like. Standard body types include, for example, the average of Japanese, the average of men (or women) in the 20s, the average of seniors, the average of children, the average of office workers, the average of specific sports athletes, etc. It is preferable to adopt and store a homologous model group. This is because a homologous model group can be configured for each purpose, and a more appropriate feature amount can be extracted by multivariate analysis of the homologous model group according to the purpose.

  The CPU 11 of the three-dimensional shape model generation apparatus 1 performs multivariate analysis based on the plurality of generated three-dimensional shape model information, and extracts a plurality of feature amounts. For example, by using principal component analysis which is a kind of multivariate analysis method, information and meanings of the entire three-dimensional shape model information group are extracted, and a distribution axis such as large pattern-small pattern is calculated. FIG. 3 is an exemplary diagram illustrating a morphological distribution situation when the modeled object is a human body. FIG. 3 shows the shape distribution status of a plurality of human bodies on two distribution axes. The distribution axis corresponds to each extracted feature amount. That is, in the example of FIG. 3, a form distribution state corresponding to two feature amounts is shown. The calculation of the morphological distribution is not limited to obtaining the relationship between the two axes, and coordinate values on more axes, that is, distribution in a multidimensional space may be calculated.

  The CPU 11 calculates the center point of the distribution axis (principal component axis) of the form of the human body, and calculates the peripheral shape by changing the value for each principal component axis. The calculated peripheral form can be specified as a coordinate value in a multidimensional space, that is, a displacement of a feature amount from the three-dimensional shape model information corresponding to the center point, and the corresponding relationship is an eigenvector matrix as an analysis result storage unit 141 stored. An example of the eigenvector matrix stored in (Equation 1) is shown.

In (Equation 1), an eigenvector matrix P for calculating n (n is a natural number) feature value T values (principal component values) based on m (m is a natural number) three-dimensional shape model information S. (Where m ≧ n). Therefore, for example, when a principal component analysis is performed as a multivariate analysis and a value of the feature amount T is given, the inverse of the eigenvector matrix P is multiplied by the inverse matrix P ⁻¹ to regenerate the three-dimensional shape model information (homology model). It can be configured.

  FIG. 4 is a flowchart showing the procedure of the 3D shape model information generation process of the CPU 11 of the 3D shape model generation apparatus 1 according to Embodiment 1 of the present invention. The CPU 11 of the three-dimensional shape model generation apparatus 1 acquires the three-dimensional shape data of the modeled object via the input unit 16 (step S401). Specifically, for example, when the object to be modeled is a human body, hundreds of thousands to millions of three-dimensional coordinate values indicating the surface shape are acquired via the input unit 16 described above.

Based on the acquired three-dimensional shape data, the CPU 11 converts the modeled object into a homologous model and estimates three-dimensional shape model information (S ₁ , S ₂ ,..., S _m ) (step S402). Then, the value of the feature amount T is calculated by multiplying the eigenvector matrix P stored in the analysis result storage unit 141 as a result of the multivariate analysis (step S403).

  Based on the value of the feature amount T, the CPU 11 estimates a position in a multidimensional feature amount space that is a multidimensional space composed of a plurality of feature amounts (step S404). In other words, even when three-dimensional shape data of a new modeled object that is not included in the population at the time of performing multivariate analysis is obtained, the feature quantity is estimated by estimating the values of multiple feature quantities. The 3D shape model information can be transformed in the space, and new 3D shape model information can be generated. In addition, the multivariate analysis population also has eigenvectors as a result of multivariate analysis for each characteristic, such as Japanese, men (or women) in the 20s, elderly people, children, office workers, specific athletes, etc. By storing the matrix in the analysis result storage unit 141, the three-dimensional shape model information according to the purpose can be reliably generated. The CPU 11 can display and output the generated three-dimensional shape model information to the display unit 17, print it to the output unit 18, and output it to an external computer via the communication unit 15. .

  Note that as the number of feature quantities in (Equation 1) increases, the more accurate three-dimensional shape model information can be reconstructed, but conversely, the calculation processing load may become excessive. Therefore, by limiting to a feature amount having a large contribution ratio among the plurality of extracted feature amounts, it is possible to reconstruct highly accurate three-dimensional shape model information while reducing the calculation processing load. The contribution rate means the ratio of the eigenvalue to the sum of all eigenvalues. For example, for a feature quantity with a contribution rate of less than 1%, the influence on the calculation result is within the range of the calculation error, and is ignored. It becomes possible to do.

  As described above, according to the first embodiment, by receiving 3D shape data composed of a large amount of actually measured data, 3D shape data of a modeling target that is not included in the homologous model group is received. Even in this case, the value of the feature amount can be specified, and the three-dimensional shape model information corresponding to the purpose can be generated based on the specified value of the feature amount.

(Embodiment 2)
Hereinafter, the three-dimensional shape model generation apparatus 1 according to Embodiment 2 of the present invention will be specifically described with reference to the drawings. Since the configuration of the three-dimensional shape model generation apparatus 1 according to the second embodiment of the present invention is the same as that of the first embodiment, detailed description is omitted by attaching the same reference numerals. The second embodiment is different from the first embodiment in that three-dimensional shape model information can be generated even when the number of actually measured data is small.

  The input means 16 is a device that acquires data (for example, dimensions, curvature, two-dimensional projection diagram, three-dimensional shape data, etc.) that can be directly measured for the modeled object. FIG. 5 is a configuration diagram illustrating an example of the input unit 16 according to the second embodiment, and the modeling target is a human body. In the example of FIG. 5, an imaging device 61 having a fixed position is provided, and the captured image data, measurement information, and the like are transmitted to the three-dimensional shape model generation device 1.

  A subject 62 as a modeling object wraps around a plurality of measuring belts 63, 63,... Each having a reflecting member for imaging at a predetermined position, or directly uses the reflecting member as a feature point on the body surface. Attach and face the imaging device 61 at the shooting position 64. After the imaging device 61 captures an image of the front of the subject 62, the subject 62 rotates 90 degrees. The imaging device 61 captures an image of the side surface of the subject 62 again.

  FIG. 6 is an exemplary view showing the configuration of the measuring belt 63. As shown in FIG. 6, a plurality of reflecting members 631, 631,... That reflect light are attached to predetermined positions of the measuring belt 63 that is wound around the human body. By calculating the distance between the predetermined positions, the perimeter can be estimated. Therefore, even with a small amount of data, it is possible to reliably acquire the data of the feature points of the human body that is the modeling target.

  FIG. 7 is a schematic diagram illustrating a display example of image data captured by the imaging device 61. As shown in FIG. 7, on the display screen, a first display area 81 for displaying image data captured from the front and a second display area 82 for displaying image data captured from the side are arranged side by side. The CPU 11 of the three-dimensional shape model generation apparatus 1 acquires actual measurement data (size, curvature, two-dimensional projection diagram, etc.) of the modeled object based on the input image data. The actual measurement data is displayed in the actual measurement value display area 83, for example.

  Note that the input unit 16 is not limited to the above-described configuration, and is a configuration capable of acquiring measured data of the modeled object and coordinate values of feature points unique to the modeled object. If there is, it will not be specifically limited.

  The display means 17 is a display device such as a CRT monitor or LCD, and displays the actual measurement data acquired by the input means 16 or the generated three-dimensional shape model. The output means 18 is a printing device such as a laser printer or a dot printer.

  Hereinafter, the operation of the three-dimensional shape model generation apparatus 1 configured as described above will be described. The three-dimensional shape model generation apparatus 1 first generates three-dimensional shape model information based on a plurality of standard three-dimensional shape data. For example, when the object to be modeled is a human body, a plurality of three-dimensional shape data of a standard body shape of the human body is homogenized, and a plurality of three-dimensional shape model information is generated as, for example, a polygon model. The standard three-dimensional shape data is composed of three-dimensional shape data measured with high accuracy by, for example, a well-known laser type three-dimensional shape measuring device, scanning device, or the like. Standard body types include, for example, the average of Japanese, the average of men (or women) in the 20s, the average of seniors, the average of children, the average of office workers, the average of specific sports athletes, etc. It is preferable to adopt and store a homologous model group. This is because a homologous model group can be configured for each purpose, and a more appropriate feature amount can be extracted by performing principal component analysis on the homologous model group corresponding to the purpose.

  The CPU 11 of the three-dimensional shape model generation apparatus 1 extracts a feature amount by performing multivariate analysis based on a plurality of generated three-dimensional shape model information. For example, by using principal component analysis which is a kind of multivariate analysis method, information and meanings of the entire three-dimensional shape model information group are extracted, and a distribution axis such as large pattern-small pattern is calculated.

  The CPU 11 calculates the center point of the distribution axis (principal component axis) of the form of the human body, and calculates the peripheral shape by changing the value for each principal component axis. The calculated peripheral form can be specified as a coordinate value in a multidimensional space, that is, a displacement of a feature amount from the three-dimensional shape model information corresponding to the center point, and the corresponding relationship is an eigenvector matrix as an analysis result storage unit 141 stored. An example of the eigenvector matrix stored in (Equation 2) is shown.

In (Expression 2), an eigenvector matrix P for calculating n (n is a natural number) feature value T (principal component value) based on m (m is a natural number) three-dimensional shape model information S. (Where m ≧ n). Therefore, for example, when a principal component analysis is performed as a multivariate analysis and a value of the feature amount T is given, the inverse of the eigenvector matrix P is multiplied by the inverse matrix P ⁻¹ to regenerate the three-dimensional shape model information (homology model). It can be configured.

  In general, the number of feature points that can be measured with a simple measuring device is from several tens to several hundred at most. This is often insufficient to reconstruct precise three-dimensional shape model information, even if the number of data points is sufficient to examine the shape distribution characteristics of the human body. Therefore, a conversion function is prepared in which the feature quantity obtained by the multivariate analysis based on the actual measurement value and the feature quantity based on the population are associated with the population of the homologous model group described above. In this way, the feature value (principal component value) based on the actual measurement value is calculated from the actual measurement value, and converted into the feature amount of the three-dimensional shape model information based on the population by the conversion function. Based on the converted feature quantity, precise three-dimensional shape model information is reconstructed.

  FIG. 8 is a flowchart showing a procedure of 3D shape model information generation processing of the CPU 11 of the 3D shape model generation apparatus 1 according to Embodiment 2 of the present invention. The CPU 11 of the three-dimensional shape model generation apparatus 1 acquires actual measurement data of the modeled object via the input unit 16 (step S801). Specifically, for example, when the modeled object is a human body, the imaging data captured by the imaging device 61 of the input unit 16 described above is acquired, and the CPU 11 determines the dimension data of the modeled object based on the captured data. To get. The actual measurement data is not limited to the dimension data, and is not particularly limited as long as it is a physical quantity that can be measured. For example, the amount of light in the imaging data of the infrared imaging device, information on the positional relationship between a plurality of feature points, information on the curvature and inflection points of an approximate curve enclosing a plurality of feature points, and the like may be used. As a result, 3D shape model information that reflects not only the information about landmarks unique to the modeled object but also the geometric features unique to the modeled object, including the relative positional relationship between landmarks, etc. Can be generated.

The CPU 11 specifies a plurality of actual measurement values U (U ₁ , U ₂ ,..., U _t ) (t is a natural number) of the modeled object based on the acquired actual measurement data (step S802). Specifically, not only the dimensional value is used as it is, but also an estimated value of the curvature of the two-dimensional projection diagram, measurement of the elongation amount of the measuring belt, and the like are included.

  The CPU 11 calculates the eigenvector matrix P ′ in advance by multivariate analysis and stores it in the analysis result storage unit 141. (Equation 3) shows an example of the eigenvector matrix P ′ calculated in advance and stored in the analysis result storage unit 141. In (Expression 3), generally, t and s are natural numbers smaller than m and n in (Expression 2).

  The CPU 11 uses the eigenvector matrix P ′ to calculate the actual measurement feature quantity T ′ related to the actual measurement value according to (Equation 3) (step S803). The CPU 11 obtains in advance a conversion function f indicating the correspondence between the actually measured feature amount T ′ and the feature amount T of the three-dimensional shape model information as shown in (Equation 4), and the three-dimensional shape model. The feature amount T of information is estimated (step S804).

  For example, when the object to be modeled is a human body, a conversion function f between an actual measurement feature amount T ′ based on the coordinate value of an anatomical feature point and the value of the feature amount T of the three-dimensional shape model information is typically represented. The feature amount T of the three-dimensional shape model information is estimated by substituting the actual measurement value feature amount T ′ calculated based on the actual measurement data into the conversion function f.

The CPU 11 multiplies the estimated plurality of feature quantities T by the inverse matrix P ⁻¹ of the eigenvector matrix P, thereby back-calculating the obtained three-dimensional shape model information S (step S805), and new three-dimensional shape model information. Is generated (step S806). The CPU 11 displays and outputs the generated three-dimensional shape model information to the display unit 17 (step S807). Of course, it may be printed out to the output means 18 or output to an external computer via the communication means 15.

  Note that the more the number of feature quantities in (Equation 2) is, the more precise 3D shape model information can be reconstructed. On the contrary, there is a possibility that the calculation processing load becomes excessive. Therefore, by limiting to a feature amount having a large contribution ratio among the plurality of extracted feature amounts, it is possible to reconstruct highly accurate three-dimensional shape model information while reducing the calculation processing load.

  As described above, according to the second embodiment, even when the number of actually measured data that can be acquired from the input unit 16 is small, a precise three-dimensional shape model corresponding to the modeled object is generated. can do. In addition, since the measurement data of the modeled object is small, for example, when generating a 3D shape model of a human body with remarkable personal characteristics, it is possible to reduce the measurement load on the subject to be modeled Become.

(Embodiment 3)
Hereinafter, the three-dimensional shape model generation apparatus 1 according to Embodiment 3 of the present invention will be specifically described with reference to the drawings. FIG. 9 is a block diagram showing a configuration of the three-dimensional shape model generation apparatus 1 according to Embodiment 3 of the present invention. Since the configuration of the three-dimensional shape model generation apparatus 1 according to the third embodiment is substantially the same as that of the second embodiment, detailed description will be given by attaching the same reference numerals to components having the same functions. Is omitted. The third embodiment includes a virtual three-dimensional shape model generation information storage unit 142 that stores the feature amount of virtual three-dimensional shape model information that has been virtually generated in advance. It is different from the first and second embodiments in that precise three-dimensional shape model information can be estimated and generated based on the information.

  Hereinafter, the operation of the three-dimensional shape model generation apparatus 1 according to the third embodiment will be described. The three-dimensional shape model generation apparatus 1 first generates three-dimensional shape model information based on standard three-dimensional shape data. For example, when the object to be modeled is a human body, a plurality of three-dimensional shape data of a standard body shape of the human body is homogenized, and a plurality of three-dimensional shape model information is generated as, for example, a polygon model. The standard three-dimensional shape data is composed of three-dimensional shape data measured with high accuracy by, for example, a well-known laser type three-dimensional shape measuring device, scanning device, or the like.

  The CPU 11 of the three-dimensional shape model generation apparatus 1 extracts feature amounts by performing multivariate analysis based on the generated three-dimensional shape model information. For example, information and meanings of the entire three-dimensional shape model information group are extracted using principal component analysis, which is a kind of multivariate analysis method, and a distribution axis such as large pattern-small pattern is calculated.

  The CPU 11 calculates the center point of the distribution axis (principal component axis) of the form of the human body, and calculates the peripheral shape by changing the value for each principal component axis. The calculated peripheral form can be specified as a coordinate value in a multidimensional space, that is, a displacement of a feature amount from the three-dimensional shape model information corresponding to the center point, and the corresponding relationship is an eigenvector matrix as an analysis result storage unit 141 stored. An example of the eigenvector matrix P is the same as (Equation 1) and (Equation 2) in the first and second embodiments.

  In general, the number of feature points that can be measured with a simple measuring device is from several tens to several hundred at most. This is often insufficient to reconstruct precise three-dimensional shape model information, even if the number of data points is sufficient to examine the shape distribution characteristics of the human body.

  Therefore, in the third embodiment, feature quantities are changed stepwise at regular intervals, virtual three-dimensional shape model information corresponding to each feature quantity value is generated in advance, for example, dimensions are measured and features are obtained. It is stored in the virtual three-dimensional shape model generation information storage unit 142 in association with the amount T. Then, based on the input actual measurement value (measurement value), necessary information for generating the adjacent virtual three-dimensional shape model information is extracted from the virtual three-dimensional shape model generation information storage unit 142, and features By interpolating the amount T, three-dimensional shape model information having the highest similarity is generated.

  FIG. 10 is a flowchart showing the procedure of the 3D shape model information generation process of the CPU 11 of the 3D shape model generation apparatus 1 according to Embodiment 3 of the present invention. The CPU 11 of the three-dimensional shape model generation apparatus 1 acquires actual measurement data of the modeled object via the input unit 16 (step S1001). Specifically, for example, when the modeled object is a human body, the imaging data captured by the imaging device 61 of the input unit 16 described above is acquired, and the CPU 11 measures the measured data of the modeled object based on the captured data. To get.

The CPU 11 specifies a plurality of actual measurement values U (U ₁ , U ₂ ,..., U _t ) (t is a natural number) of the modeled object based on the acquired actual measurement data (step S1002). Specifically, not only the dimensional value is used as it is, but also an estimated value of the curvature of the two-dimensional projection diagram, measurement of the elongation amount of the measuring belt, and the like are included.

  Based on the actual measurement value U, the CPU 11 extracts a feature amount T of a plurality of adjacent virtual three-dimensional shape model information from the three-dimensional shape model generation information storage unit 142 (step S1003), and interpolates the feature amount T. In step S1004, new three-dimensional shape model information is generated (step S1005). That is, the virtual three-dimensional shape model generation information storage unit 142 stores actual measurement values corresponding to a plurality of virtual three-dimensional shape model information in which the feature amount T is changed stepwise. Therefore, at least two pieces of virtual three-dimensional shape model information having a feature amount close to the input actual measurement value are extracted, and an average value of the feature amounts T corresponding to the extracted virtual three-dimensional shape model information is obtained. By calculating, three-dimensional shape model information approximate to the three-dimensional shape model information corresponding to the input actual measurement data can be generated.

  The CPU 11 displays and outputs the generated three-dimensional shape model information to the display unit 17 (step S1006). Of course, it may be printed out to the output means 18 or output to an external computer via the communication means 15.

  As in the second embodiment, the high-precision 3D shape model information is reproduced while reducing the processing load by limiting the extracted feature quantities to the feature quantities having a large change amount. It goes without saying that it can be configured.

  As described above, according to the third embodiment, even when only the actual measurement data for which it is difficult to specify the feature amount of the three-dimensional shape model information can be acquired, for example, the virtual three-dimensional shape described above is used. By generating a look-up table like the model storage unit 142 in advance, it is possible to generate approximately corresponding three-dimensional shape model information with high accuracy.

(Embodiment 4)
Hereinafter, a three-dimensional shape model generation system according to Embodiment 4 of the present invention will be specifically described with reference to the drawings. FIG. 11 is a block diagram showing a configuration of a three-dimensional shape model generation system according to Embodiment 4 of the present invention. The fourth embodiment includes a terminal device 4 that collects actual measurement data of a modeled object, displays the generated 3D shape model information, and the like, and a central device 5 that generates 3D shape model information. However, the present embodiment differs from the first to third embodiments in that the processing load on the terminal device 4 is greatly reduced.

  11, the three-dimensional shape model generation system according to Embodiment 4 of the present invention enables the terminal device 4 and the central device 5 to perform data communication via a network 6 such as the Internet or WAN. It is connected. The terminal device 4 connects at least a CPU (central processing unit) 41, auxiliary storage means 42, RAM 43, storage means 44, communication means 45, input means 46, display means 47, output means 48, and the hardware described above. An internal bus 49 is used.

  The CPU 41 is connected to the above-described hardware units of the terminal device 4 via the internal bus 49, and controls the above-described hardware units, and in accordance with the computer program 8 stored in the RAM 43, various types of hardware are provided. Perform software functions. The RAM 43 includes an SRAM, a flash memory, and the like, and stores a computer program (load module), temporary data generated when the computer program is executed, and the like.

  The storage unit 44 includes a built-in fixed storage device (hard disk), a ROM, and the like, and a portable recording medium 7 such as a DVD or a CD-ROM. The computer program 8 stored in the storage means 44 is downloaded by the auxiliary storage means 42 from the portable recording medium 7 such as a CD-ROM in which information such as programs and data is recorded, and from the storage means 44 to the RAM 43 at the time of execution. Deploy and execute.

  The communication means 45 is connected to an internal bus 49 and can be connected to an external network 6 such as the Internet, LAN, WAN, etc., so that data can be transmitted / received to / from an external computer such as the central device 5. It has become. Therefore, it is possible to use the three-dimensional shape model information stored in one or a plurality of external computers.

  The input means 46 is a device that acquires three-dimensional shape data of the modeled object. The input unit 46 may have the same configuration as that of the input unit 16 of the first embodiment, and can acquire actual measurement data of the modeled object and feature amount data unique to the modeled object. If it is, it will not specifically limit.

  The display means 47 is a display device such as a CRT monitor or LCD, and displays actual measurement data acquired by the input means 46 or a generated three-dimensional shape model. The output means 48 is a printing device such as a laser printer or a dot printer.

  The central device 5 includes at least a CPU (central processing unit) 51, an auxiliary storage unit 52, a RAM 53, a storage unit 54, a communication unit 55, and an internal bus 56 that connects the hardware described above.

  The CPU 51 is connected to the above-described hardware units of the central device 5 via the internal bus 56, controls the above-described hardware units, and performs various processes according to the computer program 10 stored in the RAM 53. Perform software functions. The RAM 53 includes an SRAM, a flash memory, and the like, and stores a computer program (load module), temporary data generated when the computer program is executed, and the like.

  The storage means 54 includes a built-in fixed storage device (hard disk), a ROM, and the like, and a portable recording medium 9 such as a DVD or a CD-ROM. The computer program 10 stored in the storage means 54 is downloaded by the auxiliary storage means 52 from the portable recording medium 9 such as a CD-ROM in which information such as programs and data is recorded, and from the storage means 54 to the RAM 53 at the time of execution. Deploy and execute. In addition, the storage means 54 includes an analysis result storage unit 541 that stores a result of multivariate analysis of a plurality of three-dimensional shape model information.

  When the object to be modeled is a human body, for example, three-dimensional shape model information obtained by homogenizing a plurality of three-dimensional shape data representing a Japanese body shape is generated and subjected to multivariate analysis. Note that the homogenized three-dimensional shape model information means three-dimensional shape data having the same number of landmark points and the same geometric structure and having feature points that are anatomically associated with each other. Yes. Accordingly, the number of data points, topology, etc. are the same among the three-dimensional shape data, and the data format is unified.

  In the analysis result storage unit 541, when the modeling target is a human body, for example, a plurality of three-dimensional shape model information representing a Japanese body shape is subjected to multivariate analysis, for example, principal component analysis, and a plurality of feature values (main features). Component) are extracted and stored as eigenvector matrices, average values, standard deviations, etc. constituting a plurality of feature quantities. According to the stored eigenvector matrix, average value, standard deviation, etc., it is possible to generate 3D shape model information corresponding to an arbitrary position in the multidimensional feature amount space.

  The communication means 55 is connected to an internal bus 56, and can be connected to an external network 6 such as the Internet, LAN, WAN, etc., and can exchange data with an external computer such as the terminal device 4. It has become. Therefore, it is possible to use the three-dimensional shape model information stored in one or a plurality of external computers.

  Hereinafter, the operation of the three-dimensional shape model generation system configured as described above will be described. In the three-dimensional shape model generation system, first, the central device 5 generates three-dimensional shape model information based on a plurality of standard three-dimensional shape data. For example, when the object to be modeled is a human body, a plurality of three-dimensional shape data of a standard body shape of the human body is homogenized, and a plurality of three-dimensional shape model information is generated as, for example, a polygon model. The standard three-dimensional shape data is composed of three-dimensional shape data measured with high accuracy by, for example, a well-known laser type three-dimensional shape measuring device, scanning device, or the like. Standard body types include, for example, the average of Japanese, the average of men (or women) in the 20s, the average of seniors, the average of children, the average of office workers, the average of specific sports athletes, etc. It is preferable to adopt and store a homologous model group. This is because a homologous model group can be configured for each purpose, and a more appropriate feature amount can be extracted by performing principal component analysis on the homologous model group corresponding to the purpose.

  The CPU 51 of the central device 5 performs multivariate analysis based on the generated three-dimensional shape model information and extracts a plurality of feature amounts. For example, the principal component analysis which is a kind of multivariate analysis method is used to extract information and meanings of the entire three-dimensional shape model information group, and a distribution axis such as large pattern-small pattern is calculated.

  The CPU 51 calculates the center point of the distribution axis (principal component axis) of the human body form, and calculates the peripheral form by changing the value for each principal component axis. The calculated peripheral form can be specified as a coordinate value in a multidimensional space, that is, a displacement of a feature amount from the three-dimensional shape model information corresponding to the center point, and the corresponding relationship is an eigenvector matrix as an analysis result storage unit It memorize | stores in 541. An example of the eigenvector matrix stored in (Equation 5) is shown.

In (Equation 5), an eigenvector matrix P for calculating n (n is a natural number) feature value T (principal component value) based on m (m is a natural number) three-dimensional shape model information S. (Where m ≧ n). Therefore, for example, when a principal component analysis is performed as a multivariate analysis and a value of the feature amount T is given, the inverse of the eigenvector matrix P is multiplied by the inverse matrix P ⁻¹ to regenerate the three-dimensional shape model information (homology model). It can be configured.

  In general, the number of feature points that can be measured with a simple measuring device is from several tens to several hundred at most. This is often insufficient to reconstruct precise three-dimensional shape model information, even if the number of data points is sufficient to examine the shape distribution characteristics of the human body. Therefore, a conversion function is prepared in which the feature quantity obtained by the multivariate analysis based on the actual measurement value and the feature quantity based on the population are associated with the population of the homologous model group described above. In this way, the feature value (principal component value) based on the actual measurement value is calculated from the actual measurement value, and converted into the feature amount of the three-dimensional shape model information based on the population by the conversion function. Based on the converted feature quantity, precise three-dimensional shape model information is reconstructed.

  FIG. 12 is a flowchart showing the procedure of the three-dimensional shape model information generation process of the CPU 51 of the central device 5 of the three-dimensional shape model generation system according to Embodiment 4 of the present invention. The CPU 51 of the central device 5 receives actual measurement data of the modeled object from the terminal device 4 (step S1201). Specifically, for example, when the modeled object is a human body, the terminal device 4 acquires imaging data captured by the imaging device 61 of the input unit 46 described above, and the CPU 41 of the terminal device 4 uses the imaging data. The measured data of the modeled object is acquired and transmitted to the central device 5.

The CPU 51 of the central device 5 specifies a plurality of actual measurement values U (U ₁ , U ₂ ,..., U _t ) (t is a natural number) of the modeled object based on the received actual measurement data (step S1202). ). Specifically, not only the dimensional value is used as it is, but also an estimated value of the curvature of the two-dimensional projection diagram, measurement of the elongation amount of the measuring belt, and the like are included.

  The CPU 11 calculates the eigenvector matrix P ′ in advance by multivariate analysis and stores it in the analysis result storage unit 141. (Equation 6) shows an example of the eigenvector matrix P ′ calculated in advance and stored in the analysis result storage unit 141. In (Expression 6), generally, t and s are natural numbers smaller than m and n in (Expression 2).

  The CPU 51 uses the eigenvector matrix P ′ to calculate the actual measurement feature quantity T ′ related to the actual measurement value according to (Equation 6) (step S1203). The CPU 51 obtains in advance a conversion function f indicating the correspondence between the actually measured value feature quantity T ′ and the feature quantity T of the three-dimensional shape model information as shown in (Expression 7), and the three-dimensional shape model. A feature amount T of information is estimated (step S1204).

  For example, when the object to be modeled is a human body, a conversion function f between an actual measurement feature amount T ′ based on the coordinate value of an anatomical feature point and the value of the feature amount T of the three-dimensional shape model information is typically represented. The feature amount T of the three-dimensional shape model information is estimated by substituting the actual measurement value feature amount T ′ calculated based on the actual measurement data into the conversion function f.

The CPU 51 transmits the plurality of estimated feature amounts T and eigenvector matrix P to the terminal device 4 (step S1205). The CPU 41 of the terminal device 4 receives the plurality of feature amounts T and the eigenvector matrix P estimated from the central device 5 and multiplies the received feature amount T by the inverse matrix P ⁻¹ of the received eigenvector matrix P. The coordinate value S of the obtained three-dimensional shape model information is calculated backward to generate new three-dimensional shape model information. The CPU 41 displays and outputs the generated three-dimensional shape model information to the display unit 47. Of course, it may be printed out to the output unit 48 or may be output to an external computer via the communication unit 45.

  As in the first to third embodiments, high-precision three-dimensional shape model information is reduced while reducing the processing load by limiting the extracted feature quantities to feature quantities having a large contribution rate. It goes without saying that can be reconfigured.

  Similarly to the first embodiment, the central device 5 may receive the three-dimensional shape data itself from the terminal device 4 and specify the feature value of the three-dimensional shape model information, or the third embodiment. Similarly to the above, it is difficult to specify the feature amount of the 3D shape model information based on the actual measurement data by generating a lookup table like the virtual 3D shape model storage unit 142 in advance. Even in this case, the approximately corresponding three-dimensional shape model information may be generated with high accuracy.

  As described above, according to the fourth embodiment, it is possible to reduce the processing load on the terminal device 4 that acquires measured data, displays and outputs the generated three-dimensional shape model information, and a plurality of terminal devices. Since a large amount of data can be collected from 4, 4,..., The three-dimensional shape model information can be reconstructed more precisely.

  Note that it is not preferable from the viewpoint of network load that the 3D shape model information, which is large-capacity data, flows through the network 6 every time the 3D shape model information is generated. Further, it is not preferable that the 3D shape model information itself flows through the network 6 from the viewpoints of security, personal information protection, and the like. Therefore, for example, the storage unit 44 of the terminal device 4 may include the analysis result storage unit 441, and only the eigenvector matrix that is the analysis result of the multivariate analysis may be transmitted from the central device 5 to the terminal device 4 in advance.

  By doing so, the communication load of the network 6 can be greatly reduced, and only the feature value for generating the three-dimensional shape model information on the terminal device 4 goes back and forth on the network. Even if a malicious third party attempts to intrude into the network 6, it is difficult to generate the three-dimensional shape model information, and the excellent effect of contributing to security improvement is also achieved.

  In addition, this invention is not limited to the said Example, It cannot be overemphasized that various deformation | transformation, substitution, etc. are possible if it is description in a claim.

It is a block diagram which shows the structure of the three-dimensional shape model production | generation apparatus which concerns on Embodiment 1 of this invention. It is an illustration figure which shows the feature point of a human body. It is an illustration figure which shows the form distribution situation in case a modeling target object is a human body. It is a flowchart which shows the procedure of the three-dimensional shape model production | generation process of CPU of the three-dimensional shape model production | generation apparatus which concerns on Embodiment 1 of this invention. It is a block diagram which shows an example of the input means which concerns on this Embodiment 2. It is an illustration figure which shows the structure of a measuring belt. It is a schematic diagram which shows the example of a display of the image data imaged with the imaging device. It is a flowchart which shows the procedure of the three-dimensional shape model production | generation process of CPU of the three-dimensional shape model production | generation apparatus which concerns on Embodiment 2 of this invention. It is a block diagram which shows the structure of the three-dimensional shape model production | generation apparatus which concerns on Embodiment 3 of this invention. It is a flowchart which shows the procedure of the production | generation process of the three-dimensional shape model of CPU of the three-dimensional shape model production | generation apparatus which concerns on Embodiment 3 of this invention. It is a block diagram which shows the structure of the three-dimensional shape model production | generation system which concerns on Embodiment 4 of this invention. It is a flowchart which shows the procedure of the production | generation process of the three-dimensional shape model of CPU of the central apparatus of the three-dimensional shape model production | generation system which concerns on Embodiment 4 of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Three-dimensional shape model production | generation apparatus 2, 7, 9 Portable recording medium 3, 8, 10 Computer program 4 Terminal apparatus 5 Central apparatus 6 Network network 11, 41, 51 CPU
12, 42, 52 Auxiliary storage means 13, 43, 53 RAM
14, 44, 54 Storage unit 15, 45, 55 Communication unit 16, 46 Input unit 17, 47 Display unit 18, 48 Output unit 19 Internal bus 141, 441, 541 Analysis result storage unit 142 Virtual three-dimensional shape model generation information storage Part

Claims (16)

Three-dimensional shape model information generating means for generating a group of three-dimensional shape model information based on the three-dimensional shape data of a plurality of modeled objects;
Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the 3D shape model information generated by the 3D shape model information generation means;
Analysis result storage means for storing a plurality of feature value values extracted by the feature value extraction means and three-dimensional shape model information in association with each other;
3D shape model generation means for generating 3D shape model information belonging to a predetermined group based on the value of the feature amount;
In a three-dimensional shape model generation apparatus comprising: output means for outputting generated three-dimensional shape model information;
Three-dimensional shape data receiving means for receiving input of three-dimensional shape data of a modeling target that is not subject to multivariate analysis;
Three-dimensional shape model information generating means for generating three-dimensional shape model information of the modeling object based on the three-dimensional shape data received by the three-dimensional shape data receiving means;
Means for estimating a plurality of feature value values corresponding to the three-dimensional shape model information with reference to the analysis result storage means. For the same type of modeling object, a plurality of three-dimensional shape model information that matches a predetermined condition is generated as a group of three-dimensional shape model information for each condition,
Comprising a condition receiving means for receiving information on the predetermined condition;
The feature amount extraction unit selects a corresponding group of three-dimensional shape model information based on information on the predetermined condition received by the condition reception unit, and performs a multivariate analysis on the selected group of three-dimensional shape model information. The three-dimensional shape model generation apparatus according to claim 1, wherein a plurality of feature amounts are extracted. 3D shape model information generating means for generating 3D shape model information of the modeled object based on 3D shape data of a plurality of modeled objects;
Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the 3D shape model information generated by the 3D shape model information generation means;
Analysis result storage means for storing a plurality of feature value values extracted by the feature value extraction means and three-dimensional shape model information in association with each other;
3D shape model generation means for generating 3D shape model information based on the value of the feature amount;
In a three-dimensional shape model generation apparatus comprising: output means for outputting generated three-dimensional shape model information;
Physical quantity receiving means for receiving input of information on a predetermined physical quantity that can be measured of the modeling object;
Correspondence relation calculating means for calculating a correspondence relation between the information related to the predetermined physical quantity received by the physical quantity receiving means and the values of the plurality of feature quantities extracted by the feature quantity extracting means;
The three-dimensional shape model generation means includes
Means for estimating values of a plurality of feature amounts extracted by the feature amount extraction unit based on information on the predetermined physical amount received by the physical amount reception unit and the correspondence relationship calculated by the correspondence relationship calculation unit;
A three-dimensional shape model generation apparatus comprising: means for inquiring the analysis result storage means based on a plurality of feature value values estimated by the means and generating new three-dimensional shape model information . 3D shape model information generating means for generating 3D shape model information of the modeled object based on 3D shape data of a plurality of modeled objects;
Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the 3D shape model information generated by the 3D shape model information generation means;
Analysis result storage means for storing a plurality of feature value values extracted by the feature value extraction means and three-dimensional shape model information in association with each other;
3D shape model generation means for generating 3D shape model information based on the value of the feature amount;
In a three-dimensional shape model generation apparatus comprising: output means for outputting generated three-dimensional shape model information;
Physical quantity receiving means for receiving input of information on a predetermined physical quantity that can be measured of the modeling object;
A virtual three-dimensional shape that stores virtual three-dimensional shape model information generated by changing each feature amount step by step in association with a predetermined physical amount that can be measured and a plurality of feature amount values. Model information storage means, and
The three-dimensional shape model generation means includes
Means for referring to the virtual three-dimensional shape model information storage means based on information on the predetermined physical quantity received by the physical quantity receiving means, and specifying values of a plurality of feature quantities extracted by the feature quantity extracting means; ,
Means for interpolating a plurality of feature value values specified by the means to estimate feature value values corresponding to the modeling object;
A three-dimensional shape model generation apparatus comprising: means for inquiring the analysis result storage means based on the feature value estimated by the means and generating new three-dimensional shape model information. Based on the three-dimensional shape data of the plurality of modeling objects, a group of three-dimensional shape model information is generated,
Multivariate analysis of the generated 3D shape model information to extract multiple features,
A plurality of extracted feature value values and 3D shape model information are stored in association with each other,
Generate 3D shape model information belonging to a predetermined group based on the feature value,
In the 3D shape model generation method for outputting the generated 3D shape model information,
Accepts input of 3D shape data of modeling objects that are not subject to multivariate analysis,
Based on the received 3D shape data, generate 3D shape model information of the modeled object,
A three-dimensional shape model generation method characterized by estimating a plurality of feature value values corresponding to three-dimensional shape model information. A plurality of three-dimensional shape model information that matches a predetermined condition is generated as a group of three-dimensional shape model information according to a condition for the same type of modeling object,
Receiving information on the predetermined condition;
Selecting a corresponding group of 3D shape model information based on the received information on the predetermined condition, and extracting a plurality of feature amounts by multivariate analysis of the selected group of 3D shape model information The method for generating a three-dimensional shape model according to claim 5. Based on the three-dimensional shape data of a plurality of modeling objects, three-dimensional shape model information of the modeling object is generated,
Multivariate analysis of the generated 3D shape model information to extract multiple features,
A plurality of extracted feature value values and 3D shape model information are stored in association with each other,
Generate 3D shape model information based on the feature value,
In the 3D shape model generation method for outputting the generated 3D shape model information,
Accepts input of information about a predetermined physical quantity that can be measured for the modeled object,
Calculate the correspondence between the received information on the predetermined physical quantity and the values of the extracted feature quantities,
Based on the received information on the predetermined physical quantity and the calculated correspondence relationship, estimate the values of the extracted feature quantities,
A three-dimensional shape model generation method characterized by generating new three-dimensional shape model information based on a plurality of estimated feature value values. Based on the three-dimensional shape data of a plurality of modeling objects, three-dimensional shape model information of the modeling object is generated,
Multivariate analysis of the generated 3D shape model information to extract multiple features,
A plurality of extracted feature value values and 3D shape model information are stored in association with each other,
Generate 3D shape model information based on the feature value,
In the 3D shape model generation method for outputting the generated 3D shape model information,
Accepts input of information about a predetermined physical quantity that can be measured for the modeled object,
The virtual three-dimensional shape model information generated by changing each feature amount stepwise is stored in association with a predetermined physical amount that can be measured and a plurality of feature amount values,
Based on the received information on the predetermined physical quantity, identify the extracted values of multiple feature quantities,
Interpolate multiple identified feature values to estimate feature values corresponding to the modeled object,
A three-dimensional shape model generation method characterized by generating new three-dimensional shape model information based on an estimated feature value. Computer
Three-dimensional shape model information generating means for generating a group of three-dimensional shape model information based on three-dimensional shape data of a plurality of modeled objects;
Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the three-dimensional shape model information generated by the three-dimensional shape model information generation means;
Analysis result storage means for storing a plurality of feature quantity values extracted by the feature quantity extraction means and the three-dimensional shape model information in association with each other;
It can be executed by a computer that functions as three-dimensional shape model generation means for generating three-dimensional shape model information belonging to a predetermined group based on the value of the feature quantity, and as output means for outputting the generated three-dimensional shape model information. In a possible computer program,
The computer,
3D shape data receiving means for receiving input of 3D shape data of a modeling object that is not subject to multivariate analysis;
Based on the three-dimensional shape data received by the three-dimensional shape data receiving means, refer to the three-dimensional shape model information generating means for generating the three-dimensional shape model information of the modeled object, and the analysis result storage means A computer program that functions as means for estimating a plurality of feature value values corresponding to three-dimensional shape model information. The computer,
Means for generating a plurality of three-dimensional shape model information that matches a predetermined condition as a group of three-dimensional shape model information according to a condition for the same type of modeling object; and a condition receiving means for receiving information on the predetermined condition Function as
The feature quantity extraction means is
A corresponding group of 3D shape model information is selected based on information on the predetermined condition received by the condition receiving means, and a plurality of feature amounts are extracted by multivariate analysis of the selected group of 3D shape model information. The computer program according to claim 9, wherein the computer program is made to function as a means for performing the operation. Computer
3D shape model information generating means for generating 3D shape model information of the modeled object based on 3D shape data of a plurality of modeled objects;
Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the three-dimensional shape model information generated by the three-dimensional shape model information generation means;
Analysis result storage means for storing a plurality of feature quantity values extracted by the feature quantity extraction means and the three-dimensional shape model information in association with each other;
In a computer program that can be executed by a computer that functions as a three-dimensional shape model generation unit that generates three-dimensional shape model information based on a feature value, and an output unit that outputs the generated three-dimensional shape model information ,
The computer,
A physical quantity acceptance means for accepting input of information relating to a predetermined physical quantity that can be measured of the modeling object;
Correspondence calculation means for calculating the correspondence between the information about the predetermined physical quantity received by the physical quantity reception means and the values of the plurality of feature quantities extracted by the feature quantity extraction means;
Means for estimating the values of a plurality of feature quantities extracted by the feature quantity extraction means based on the information related to the predetermined physical quantity received by the physical quantity acceptance means and the correspondence relation calculated by the correspondence relation calculation means; and A computer program for inquiring of the analysis result storage means based on a plurality of feature value values estimated by the means and generating new three-dimensional shape model information. Computer
3D shape model information generating means for generating 3D shape model information of the modeled object based on 3D shape data of a plurality of modeled objects;
Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the three-dimensional shape model information generated by the three-dimensional shape model information generation means;
Analysis result storage means for storing a plurality of feature quantity values extracted by the feature quantity extraction means and the three-dimensional shape model information in association with each other;
In a computer program that can be executed by a computer that functions as a three-dimensional shape model generation unit that generates three-dimensional shape model information based on a feature value, and an output unit that outputs the generated three-dimensional shape model information ,
The computer,
A physical quantity acceptance means for accepting input of information relating to a predetermined physical quantity that can be measured of the modeling object;
A virtual three-dimensional shape that stores virtual three-dimensional shape model information generated by changing each feature amount step by step in association with a predetermined physical amount that can be measured and a plurality of feature amount values. Model information storage means,
Means for referring to the virtual three-dimensional shape model information storage means based on the information on the predetermined physical quantity received by the physical quantity receiving means, and specifying values of a plurality of feature quantities extracted by the feature quantity extraction means;
Means for interpolating a plurality of feature value values specified by the means to estimate a feature value corresponding to the modeled object, and the analysis result based on the feature value estimated by the means A computer program which functions as a means for inquiring storage means and generating new three-dimensional shape model information. In a three-dimensional shape model generation system having a central device connected to be capable of data communication via a network and a terminal device,
The central device is
Three-dimensional shape model information generating means for generating a group of three-dimensional shape model information based on the three-dimensional shape data of a plurality of modeled objects;
Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the 3D shape model information generated by the 3D shape model information generation means;
Analysis result storage means for storing information on the correspondence between the values of the plurality of feature quantities extracted by the feature quantity extraction means and the three-dimensional shape model information;
3D shape data receiving means for receiving 3D shape data of a modeled object that is not subject to multivariate analysis;
Three-dimensional shape model information generating means for generating three-dimensional shape model information of the modeling object based on the three-dimensional shape data received by the three-dimensional shape data receiving means;
Means for estimating a plurality of feature value corresponding to the three-dimensional shape model information with reference to the analysis result storage means;
A feature amount transmitting means for transmitting a plurality of feature amount values estimated by the means and information relating to correspondence stored in the analysis result storage means to the terminal device;
The terminal device
3D shape data acquisition means for acquiring 3D shape data of a modeled object that is not subject to multivariate analysis;
Three-dimensional shape data transmitting means for transmitting the acquired three-dimensional shape data to the central device;
Receiving means for receiving information about the estimated values of the plurality of feature quantities and the correspondence relationship from the central device;
Means for generating three-dimensional shape model information based on the received information about the values of the plurality of feature values and the correspondence relationship;
A three-dimensional shape model generation system comprising: output means for outputting a plurality of feature value values and reconstructed three-dimensional shape model information. The central device is
For the same type of modeling object, a plurality of three-dimensional shape model information that matches a predetermined condition is generated as a group of three-dimensional shape model information for each condition,
The terminal device
Condition accepting means for accepting information on the predetermined condition;
Means for transmitting information relating to the predetermined condition received by the condition receiving means to the central device,
The central device is
Means for receiving information on the predetermined condition from the terminal device;
The feature amount extraction unit selects a corresponding group of three-dimensional shape model information based on the information on the predetermined condition received by the unit, and performs a multivariate analysis on the selected group of three-dimensional shape model information. The three-dimensional shape model generation system according to claim 13, wherein the feature amount is extracted. In a three-dimensional shape model generation system having a central device connected to be capable of data communication via a network and a terminal device,
The central device is
3D shape model information generating means for generating 3D shape model information of the modeled object based on 3D shape data of a plurality of modeled objects;
Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the 3D shape model information generated by the 3D shape model information generation means;
Analysis result storage means for storing information on the correspondence between the values of the plurality of feature quantities extracted by the feature quantity extraction means and the three-dimensional shape model information;
Physical quantity receiving means for receiving information on a predetermined physical quantity that can be measured of the modeling object;
Correspondence calculation means for calculating a correspondence relationship between the information related to the predetermined physical quantity received by the physical quantity receiving means and the values of the plurality of feature quantities extracted by the feature quantity extraction means;
Means for estimating values of a plurality of feature amounts extracted by the feature amount extraction unit based on information on the predetermined physical amount received by the physical amount reception unit and the correspondence relationship calculated by the correspondence relationship calculation unit;
A feature amount transmitting means for transmitting a plurality of feature amount values estimated by the means and information relating to correspondence stored in the analysis result storage means to the terminal device;
The terminal device
Physical quantity acquisition means for acquiring information on a predetermined physical quantity that can be measured by the modeling target;
Physical quantity transmitting means for transmitting information on the acquired physical quantity to the central device;
Receiving means for receiving information about the estimated values of the plurality of feature quantities and the correspondence relationship from the central device;
Means for generating three-dimensional shape model information based on the received information about the values of the plurality of feature values and the correspondence relationship;
A three-dimensional shape model generation system comprising: output means for outputting a plurality of feature value values and reconstructed three-dimensional shape model information. In a three-dimensional shape model generation system having a central device connected to be capable of data communication via a network and a terminal device,
The central device is
3D shape model information generating means for generating 3D shape model information of the modeled object based on 3D shape data of a plurality of modeled objects;
Feature quantity extraction means for extracting a plurality of feature quantities by multivariate analysis of the 3D shape model information generated by the 3D shape model information generation means;
Analysis result storage means for storing information on the correspondence between the values of the plurality of feature quantities extracted by the feature quantity extraction means and the three-dimensional shape model information;
Physical quantity receiving means for receiving information on a predetermined physical quantity that can be measured of the modeling object;
A virtual three-dimensional shape that stores virtual three-dimensional shape model information generated by changing each feature amount step by step in association with a predetermined physical amount that can be measured and a plurality of feature amount values. Model information storage means;
Means for referring to the virtual three-dimensional shape model information storage means based on information on the predetermined physical quantity received by the physical quantity receiving means, and specifying values of a plurality of feature quantities extracted by the feature quantity extracting means; ,
Means for interpolating a plurality of feature value values specified by the means to estimate feature value values corresponding to the modeling object;
A feature amount transmitting means for transmitting a plurality of feature amount values estimated by the means and information relating to correspondence stored in the analysis result storage means to the terminal device;
The terminal device
Physical quantity acquisition means for acquiring information on a predetermined physical quantity that can be measured by the modeling target;
Physical quantity transmitting means for transmitting information on the acquired physical quantity to the central device;
Receiving means for receiving information about the estimated values of the plurality of feature quantities and the correspondence relationship from the central device;
Means for generating three-dimensional shape model information based on the received information about the values of the plurality of feature values and the correspondence relationship;
A three-dimensional shape model generation system comprising: output means for outputting a plurality of feature value values and reconstructed three-dimensional shape model information.

Images