Patent Application
20240358277
This application is based upon and claims the benefit of priority from the prior Japanese Patent Application No. 2023-073079 filed on Apr. 27, 2023, the entire contents of which are incorporated herein by reference.
The present disclosure relates to a body shape analysis device and a body shape analysis method.
It is preferable to prepare many types of shoe size variations for various foot shapes and sizes. However, since increasing the size variations entails an increase in manufacturing costs and the risk of holding a large stock, many shoe manufacturers mainly prepare size variations that are in high demand.
With regard to the shoe size description, since the unit and standards of size vary depending on the country or region, when shoes of the same size are sold in different countries or regions, the size description needs to be adjusted for each country or region. Accordingly, it is difficult to strictly conform the size description of shoes sold internationally to the size description in every country or region. Therefore, it can be sometimes difficult to know which size fits your feet without actually trying the shoes on. Some shoe stores can be staffed with a professional called a shoe fitter who can select shoes that fit a customer's feet, based on a wealth of knowledge and experience. However, online sales cannot rely on shoe fitters. Meanwhile, in recent years, there has been developed a technology of measuring foot shapes three-dimensionally to select shoes that fit the feet.
There is a known technology of selecting a type of shoe that fits a foot shape, based on three-dimensional data obtained by measuring the foot shape three-dimensionally (see Japanese Unexamined Patent Application Publication No. 2000-90272, for example). In Japanese Unexamined Patent Application Publication No. 2000-90272, from the three-dimensional foot shape data, feature data of a two-dimensional shape is generated for each of the toe portion, central portion, and heel portion. Similarly, feature data for each portion is also generated for lasts. The feature data of a large number of lasts are collected and clustered, and each cluster is stored with a code assigned thereto. The relationship between the feature data of a foot shape and the feature data of a last that fits the foot shape is learned with a neural network, and a rule is defined such that the feature data of a last is obtained from the feature data of a foot shape.
In general, commercially available shoes are often made separately for men and women, taking into consideration differences in physical size and design preferences between men and women. In Japan, in particular, size variations are often provided separately for men and women in accordance with the Japanese Industrial Standards (JIS), so that shoes are rarely designed as unisex models. Except for sports shoes for which the fit is particularly important, size variations based on the foot width, besides the foot length, are rarely prepared.
When shoes are selected based on the foot length, a mismatch can occur such that the foot width is too narrow or too wide. Also, there can be a case where, because of lack of fit in a part other than the foot length or foot width, shoes that fit the feet cannot be found. However, since size variations are basically provided based on the foot length, which is the factor having the greatest impact on the fit of shoes, it would be complicated to increase the shoes to be considered for purchase by adding shoes of different foot length sizes, even though the fit in the foot width or other parts is considered important. Especially, in online sales, trying multiple sizes is not easy.
Meanwhile, for shoe manufacturers, since there are differences in demand for sizes and shapes due to differences in average physical size among countries or regions, it is particularly difficult to prepare size variations that meet the demand for shoes to be sold internationally.
The present disclosure has been made in view of such a situation, and a general purpose thereof is to provide a technology for improving the fit between a foot and a shoe.
In response to the above issue, a body shape analysis device according to one embodiment of the present disclosure includes: a model storage unit that stores, in advance, a plurality of three-dimensional homologous models for a plurality of body shapes with a three-dimensional coordinate group indicating an anatomical feature of at least part of a body; a measurement data acquirer that acquires, as measurement data, a three-dimensional coordinate group indicating an anatomical feature corresponding to the three-dimensional homologous model of a body shape of a measurement subject; and an evaluation determination unit that determines, among the plurality of three-dimensional homologous models, a three-dimensional homologous model similar to a body shape of the measurement subject by evaluating similarity in coordinates between a three-dimensional coordinate group indicating an anatomical feature in the measurement data and a three-dimensional coordinate group indicating an anatomical feature in each of the plurality of three-dimensional homologous models.
Another embodiment of the present disclosure also relates to a body shape analysis device. The device includes: a model storage unit that stores, in advance, a plurality of three-dimensional homologous models for a plurality of body shapes with a three-dimensional coordinate group indicating an anatomical feature of a foot, in which the plurality of three-dimensional homologous models are classified into a plurality of clusters by cluster analysis based on cosine similarity between position vectors from a predetermined reference point to the respective points in the three-dimensional homologous models; and an output unit that outputs a three-dimensional coordinate group of a three-dimensional homologous model in each of the clusters.
Yet another embodiment of the present disclosure relates to a body shape analysis method. The method includes: reading a plurality of three-dimensional homologous models for a plurality of body shapes with a three-dimensional coordinate group indicating an anatomical feature of a foot; acquiring, as measurement data, a three-dimensional coordinate group indicating an anatomical feature corresponding to the three-dimensional homologous model of a body shape of a measurement subject; evaluating similarity in coordinates between a three-dimensional coordinate group indicating an anatomical feature in the measurement data and a three-dimensional coordinate group indicating an anatomical feature in each of the plurality of three-dimensional homologous models; and determining, among the plurality of three-dimensional homologous models, a three-dimensional homologous model similar to a body shape of the measurement subject, based on evaluation for the similarity.
Embodiments will now be described, by way of example only, with reference to the accompanying drawings which are meant to be exemplary, not limiting, and wherein like elements are numbered alike in several Figures, in which:
In the following, the present disclosure will be described based on preferred embodiments with reference to each drawing. In the embodiments and modifications, like reference characters denote like or corresponding constituting elements, and the repetitive description will be omitted as appropriate.
In the present embodiment, three-dimensional shapes of feet of a large number of measurement subjects are measured and accumulated as foot shape data, which are then classified into multiple clusters using a cluster analysis method described later. Also, by designing lasts or sock liners based on the classified multiple clusters, shape variations of shoe products can be provided based on standards different from conventional standards, and shoes that fit many people's feet can be provided. Further, demand for each shape in the shape variations is forecasted based on a foot shape distribution.
The three-dimensional foot shape measuring device 18 acquires three-dimensional data related to a foot shape of the measurement subject 10 by laser measurement. The “foot shape” as used herein is a three-dimensional model that virtually reproduces a three-dimensional shape of a foot of the measurement subject 10. The measurement values as the result of three-dimensional measurement of a foot shape scanned by the three-dimensional foot shape measuring device 18 are transmitted from the three-dimensional foot shape measuring device 18 to the shape analysis device 50. The shape analysis device 50 can be implemented as a server to which multiple three-dimensional foot shape measuring devices 18 connect via a network and that collects and analyzes measurement data from each of the multiple three-dimensional foot shape measuring devices 18. The “body shape analysis device” in the claims can mean the entire shape analysis system 100 or can mean the shape analysis device 50. Also, in the present embodiment, the shape analysis device 50 substantially corresponds to the “body shape analysis device” because many of the characteristic functions included in the “body shape analysis device” in the claims are implemented such as to be included in the shape analysis device 50.
The shape analysis device 50 is a server computer connected to multiple three-dimensional foot shape measuring devices 18 via a network line, such as the Internet or a local area network (LAN), and a communication means, such as wireless communication. The shape analysis device 50 can be constituted by a single server computer or can be constituted by a combination of multiple server computers.
The model storage unit 22 stores, in advance, multiple three-dimensional models for multiple body shapes with a three-dimensional coordinate group indicating anatomical features of at least part of a body. The body shape as used herein mainly means a foot shape but can also mean a body shape other than a foot shape, such as physical type data indicating the shape of the entire human body, or a partial body shape other than a foot shape. The three-dimensional model of a foot shape stored in the model storage unit 22 is, for example, a three-dimensional homologous model in which a coordinate group of points indicating anatomical features of an average foot in three-dimensional space is defined. A foot shape three-dimensional homologous model is a model with which multiple foot shapes having various sizes and shapes are each represented by a polyhedron anatomically defined by the same topological structure with the same number of points. The points representing anatomical features of an average foot are obtained in advance based on foot samples obtained from a large number of subjects. The three-dimensional models in modifications can be various other models created in accordance with the same anatomical criteria, such as a three-dimensional shape model in a spherical coordinate system, and a model obtained by modifying a template foot shape generated by a three-dimensional CAD system, based on anatomical landmarks.
Referring back to
The analysis processing unit 30 includes a cluster analysis unit 31 and a principal component analysis unit 32. The cluster analysis unit 31 performs cluster analysis for an upper layer and a lower layer, which will be described later, on multiple three-dimensional homologous models stored in the model storage unit 22. The cluster analysis unit 31 performs a method in which different cluster analysis methods for the upper layer and lower layer are combined. The principal component analysis unit 32 performs principal component analysis, which will be described later, on multiple foot shape three-dimensional homologous models stored in the model storage unit 22. After performing the cluster analysis for the upper layer, the cluster analysis unit 31 performs the cluster analysis for the lower layer based on the result of the principal component analysis performed by the principal component analysis unit 32.
As the cluster analysis unit 31 performs the cluster analysis and the principal component analysis unit 32 performs the principal component analysis, information of each cluster is acquired as the classification results and output to the outside by the output unit 26. Based on the classification results of the cluster analysis, the analysis processing unit 30 divides the three-dimensional homologous models stored in the model storage unit 22 into multiple clusters.
The output unit 26 can display the classification results on a display screen of the shape analysis device 50 or an external terminal 17 or can transmit the classification results to an external terminal 17 via the communication unit 19. The output unit 26 outputs three-dimensional coordinate groups of three-dimensional homologous models of each cluster. The information of the classification results output by the output unit 26 can be used for designing of lasts or sock liners.
The statistical processing unit 35 performs demand forecasting for a last corresponding to each foot shape model based on the foot shape model distribution for each cluster. The result of the demand forecasting is output to the external terminal 17 by the output unit 26. Based on the result of the demand forecasting, the statistical processing unit 35 manages a database of lasts, footwear, or sock liners corresponding to the foot shape models in each cluster stored in the model storage unit 22.
The multiple diagonal lines ascending to the right in
The distribution of the women's samples shown in
The bold diagonal line in each of
Therefore, the cluster analysis unit 31 in the present embodiment performs cluster analysis across all foot length sizes, as shown in
The Euclidean distance (distance 124) from the first point 120 to the second point 121 is shorter than the Euclidean distance (distance 125) from the first point 120 to the third point 122. Therefore, the second point 121 is judged closer to the first point 120 than the third point 122 based on the Euclidean distance.
On the other hand, the angle θ2 between the position vectors of the first point 120 and the third point 122 from the origin is smaller than the angle θ1 between the position vectors of the first point 120 and the second point 121 from the origin, and the cosine value of the angle θ2 is closer to 1. Therefore, the third point 122 is judged closer to the first point 120 than the second point 121 based on the cosine similarity.
Further, the first point 120 and the fourth point 123 have almost the same position vectors from the origin and hence the angle θ is almost zero, so that the cosine similarity is almost 1. Therefore, the fourth point 123 is farther from the first point 120 than the second point 121 based on the Euclidean distance but is infinitely closer to the first point 120 based on the cosine similarity.
In the example of
In the example of
In the cluster analysis based on the Euclidean distance, the foot length gradually increases from left to right in the subdivided clusters in the lowest layer, with the foot length size being smaller toward the left and being larger toward the right. This indicates that the influence of the foot length size is dominant in the cluster analysis based on the Euclidean distance. Also, in terms of sex, males gradually increases from left to right, with females increasing toward the left and males increasing toward the right. This is because women tend to have smaller foot lengths than men, and it can be seen, as a result, that the cluster analysis based on the Euclidean distance is also affected by sex.
Meanwhile, in the cluster analysis based on the cosine similarity, various foot lengths are randomly mixed from left to right in the subdivided clusters in the lowest layer, indicating that there is little influence by the foot length size. Similarly, sex is also randomly mixed, indicating that there is also little influence by sex.
The top three clusters (a first cluster 150, a second cluster 151, and a third cluster 152) are regarded as the upper layer. In each of the first cluster 150, second cluster 151, and third cluster 152, a foot shape based on the coordinates of each point obtained by normalizing the foot length size is defined as an average foot.
By superimposing the foot shape model of the average foot of each cluster, the characteristics of each cluster and the areas where such characteristics are likely to appear can be determined based on differences in partial shapes. For example, a characteristic can be found in a difference in instep height, such as, in a direction of instep height, the instep of the first cluster 150 is the lowest, the instep of the second cluster 151 is intermediate, and the instep of the third cluster 152 is the highest. For example, a characteristic can be found in a difference in first toe angle, such as, with regard to the angle of the first toe, the angle of the first cluster 150 extends toward the medial side the most, the angle of the second cluster 151 is intermediate, and the angle of the third cluster 152 is slightly inclined toward the second toe. For example, a characteristic can be found in the inclination around the ankle, such as, with regard to the inclination around the ankle, the inclination of the first cluster 150 is the largest on the medial side, the inclination of the second cluster 151 is intermediate, and the inclination of the third cluster 152 is the smallest.
The principal component analysis unit 32 performs principal component analysis on the foot shape data in each cluster, calculates data obtained by adding a first principal component PC1, as the variance, to the average foot that is the center of gravity of each cluster, and plots the relationship between the foot length and the ball girth in the data on the graph of
The clusters have been obtained by primary clustering based on the cosine similarity and classified such that the influence of the foot length size can be ignored, but the data itself contains foot length information. Therefore, when the principal component analysis unit 32 performs the principal component analysis, a principal component greatly affected by the foot length is extracted as the first principal component PC1. However, since the data have been classified based on the cosine similarity into clusters having different average foot shapes as the large classifications in the upper layer, the classification based on principal components in the clustering for each cluster in the lower layer can be performed easily. In this regard, the plots with open circles in the graph, which indicate the relationship between the foot length and the ball girth in the data of the average foot+the first principal component PC1 for each cluster, are located almost along the regression line of each cluster, indicating that the results of the cluster analysis and the results of the principal component analysis are consistent.
As shown in the two upper graphs on the left and right sides, the first principal component PC1 is strongly correlated not only with the foot length but also with the ball girth. On the other hand, as shown in the middle and lower graphs, the regression lines therein are almost horizontal, so that it can be seen that the second principal component PC2 and the subsequent principal component are hardly affected by the foot length or ball girth. Utilizing such characteristics in the results of the principal component analysis, in the present embodiment, the first principal component PC1, which is affected by the foot length and ball girth, is excluded, and only the second principal component PC2 and the subsequent principal components, which are not affected by the foot length or ball girth, are used for sub-clustering. The utilization of such characteristics is considered to be convenient for construction of variations based on the clustering focusing on differences in shape of a detail, which are different from the conventional grading based on the size.
As shown in
A group of foot shape models classified into a large number of clusters based on detailed features can be used not only for designing of last and footwear variations, but also for designing of sock liners. For example, variations associated with the shape of a detail, which does not depend on the foot length and ball girth sizes, can be absorbed by design variations of sock liners, thereby reducing the number of variations to be absorbed in the designing of lasts and footwear. That is, by combining the last and footwear variations and the sock liner variations, the overall variations can be designed to cover the entire foot shape model group.
Also, a large number of foot shape models included in the foot shape model group can be allocated to size variations according to the JIS, based on which of the lasts of size variations designed according to the JIS the foot shape model fits.
Since the number of people of the foot shape data included in each cluster is different, the ratio of the number of people in each cluster to the entire foot shape data group, as the population, is also different. Meanwhile, when a group of foot shape data collected and accumulated for various purposes are held as a database besides the foot shape data group used for the cluster analysis in the first embodiment, the distribution for each foot length size with respect to the population can be calculated based on those data. Therefore, the statistical processing unit 35 can quantify the ratio of each combination of a cluster and a foot length by combining the ratio of the number of people in each cluster and the distribution for each foot length size. The ratio of each combination of a cluster and a foot length when the foot length sizes are divided in the range of 22 cm to 31 cm in increments of 1 cm is quantified.
When a variation of a combination pattern for the size and shape of a footwear product or sock liner product is designed based on the ratio of each combination of a cluster and a foot length, the statistical processing unit 35 can forecast demand for initial stock according to the ratio of each combination of a cluster and a foot length. In such a case, a combination pattern of a cluster and a foot length for which demand is expected to be extremely low can be excluded from the shoe product variations from the beginning. Besides forecasting demand for initial stock, the statistical processing unit 35 can also reflect a demand forecast in the production plan when the stock decreases due to actual sales. The statistical processing unit 35 can output information regarding the ratio of each combination of a cluster and a foot length, to the outside via the output unit 26.
The present embodiment differs from the first embodiment in that a three-dimensional shape of a foot of a measurement subject is measured to acquire foot shape data, and a foot shape similar thereto can be extracted from foot shape three-dimensional homologous models accumulated in the shape analysis device 50 and can be used as reference information for purchase of shoes. In the following, description will be given mainly for the differences from the first embodiment, and the explanation of features in common will be omitted.
The information terminal 16 can generate a foot shape of a measurement subject 10 by means of a three-dimensional scanner function using a technology such as light detection and ranging (LiDAR) or through image synthesis processing such as photogrammetry. The shape analysis system 100 can further include at least one of the measurement mat 12, the three-dimensional foot shape measuring device 18, or the information terminal 16. The information terminal 16 can be operated by the measurement subject 10 himself or herself as a user, or the three-dimensional foot shape measuring device 18 or the information terminal 16 can be used by a person other than the measurement subject 10, such as a salesperson of a shoe store, to perform the foot shape scanning. Therefore, the information terminal 16 can be a terminal owned by a shoe store where the three-dimensional foot shape measuring device 18 is installed or can be a terminal owned by the measurement subject 10 himself or herself.
The measurement values as the results of three-dimensional measurement of a foot shape scanned by the three-dimensional foot shape measuring device 18 or the information terminal 16 are transmitted from the three-dimensional foot shape measuring device 18 or the information terminal 16 to the shape analysis device 50. The shape analysis device 50 can be implemented as a server to which multiple three-dimensional foot shape measuring devices 18 or information terminals 16 connect via a network and that collects and analyzes measurement data from each of the three-dimensional foot shape measuring devices 18 or information terminals 16. The “body shape analysis device” in the claims can mean the entire shape analysis system 100 or can mean the shape analysis device 50. Also, in the present embodiment, the shape analysis device 50 substantially corresponds to the “body shape analysis device” because many of the characteristic functions included in the “body shape analysis device” in the claims are implemented such as to be included in the shape analysis device 50. The characteristic functions of the “body shape analysis device” can also be distributed between the information terminal 16 and the shape analysis device 50 or can also be implemented such that many of them are included in the information terminal 16.
The shape analysis device 50 is a server computer connected to multiple three-dimensional foot shape measuring devices 18 or multiple information terminals 16 via a network line, such as the Internet or a LAN, and a communication means, such as wireless communication. The information terminal 16 can be a mobile information terminal, such as a smartphone or a tablet terminal, or can be a personal computer.
Multiple foot shape three-dimensional homologous models to be stored in the model storage unit 22 can be classified into multiple clusters by the cluster analysis described in the first embodiment. Alternatively, the models can be three-dimensional homologous models that respectively correspond to last, footwear, or sock liner variations designed based on the classification results of the cluster analysis of the first embodiment. Alternatively, the models can be conventional three-dimensional homologous models that each correspond to a combination of a foot length and a ball girth in accordance with the JIS.
The measurement data acquirer 20 acquires, as measurement data, a three-dimensional coordinate group indicating anatomical features corresponding to a three-dimensional homologous model of the foot shape of the measurement subject 10, from the three-dimensional foot shape measuring device 18 or the information terminal 16 via the communication unit 19. The model generator 21 generates a foot shape three-dimensional homologous model based on the measurement data acquired by the measurement data acquirer 20 and outputs the model as a foot shape three-dimensional homologous model of the measurement subject 10 to the evaluation determination unit 24.
The evaluation determination unit 24 evaluates the similarity in coordinates between the three-dimensional coordinate group indicating anatomical features in the measurement data of the measurement subject 10 and the three-dimensional coordinate groups indicating anatomical features in multiple three-dimensional homologous models stored in the model storage unit 22. Based on the similarity in coordinates between corresponding three-dimensional homologous models, the evaluation determination unit 24 determines, among the multiple three-dimensional homologous models, a three-dimensional homologous model similar to the foot shape of the measurement subject 10.
More specifically, the evaluation determination unit 24 acquires a position vector from a predetermined reference point to each point in the multiple three-dimensional homologous models and also acquires a position vector from a predetermined reference point to each point in the three-dimensional homologous model in the measurement data. Based on the distance between points or the similarity between position vectors in corresponding three-dimensional homologous models, the evaluation determination unit 24 selects, from among the multiple three-dimensional homologous models, a three-dimensional homologous model similar to the foot shape of the measurement subject 10. The distance between points can be the Euclidean distance, and the similarity between position vectors can be the cosine similarity, for example. The evaluation determination unit 24 combines the selection result for a similar three-dimensional homologous model based on the Euclidean distance and the selection result for a similar three-dimensional homologous model based on the cosine similarity and determines therebetween the three-dimensional homologous model that is most similar to the foot shape of the measurement subject 10. It is assumed here that about 6,000 foot shape three-dimensional homologous models are stored in the model storage unit 22. In such a case, even if the evaluation determination unit 24 determines the most similar three-dimensional homologous model by calculating the similarity with all the three-dimensional homologous models, it can be determined in about 10 seconds of calculation, so that a foot shape can be extracted in a very short time.
The evaluation determination unit 24 scales the three-dimensional homologous model thus determined to various sizes based on the foot length of the measurement subject 10 and calculates the similarity between each of the scaled models and the foot shape three-dimensional homologous model of the measurement subject 10. Based on the similarity of each scaled model, the evaluation determination unit 24 determines the scale of the three-dimensional homologous model with the highest similarity. The evaluation determination unit 24 extracts, from the database in the model storage unit 22, the determined foot shape three-dimensional homologous model and the type of shoe or sock liner corresponding to the scale and generates information to be presented to the measurement subject 10.
The output unit 26 transmits, to the information terminal 16 via the communication unit 19, information on a foot shape model similar to a foot of the measurement subject 10 and recommendation information indicating the type of shoe or sock liner corresponding to the foot shape, determined by the evaluation determination unit 24.
In a modification, the evaluation determination unit 24 can evaluate which of the foot shape three-dimensional homologous models of multiple celebrities stored in the model storage unit 22 has the highest similarity, and the evaluation result can be output to the information terminal 16.
The present disclosure has been described based on embodiments. The embodiments are intended to be illustrative only, and it will be obvious to those skilled in the art that various modifications to a combination of constituting elements or processes in the embodiments could be developed and that such modifications also fall within the scope of the present disclosure. Also, when the embodiments set forth above are generalized, the following aspects are obtained.
A body shape analysis device, including:
The body shape analysis device according to Aspect 1, wherein the evaluation determination unit acquires a position vector from a predetermined reference point to each point in the plurality of three-dimensional homologous models, also acquires a position vector from the predetermined reference point to each point in a three-dimensional homologous model in the measurement data, and determines, based on the distance between points or the similarity between position vectors in corresponding three-dimensional homologous models, a three-dimensional homologous model similar to a body shape of the measurement subject, among the plurality of three-dimensional homologous models.
The body shape analysis device according to Aspect 1 or 2, wherein the plurality of three-dimensional homologous models are classified into a plurality of clusters by cluster analysis based on cosine similarity between position vectors from a predetermined reference point to the respective points in the three-dimensional homologous models and are stored in the model storage unit.
A body shape analysis device, including:
The body shape analysis device according to any one of Aspects 1 through 4, wherein the plurality of three-dimensional homologous models are classified into a plurality of clusters in an upper layer by cluster analysis based on cosine similarity and are further classified into a plurality of more detailed clusters in a lower layer by performing cluster analysis on each cluster in the upper layer based on principal component analysis of which results are dimensionally reduced to the second and the subsequent principal components excluding the first principal component.
The body shape analysis device according to any one of Aspects 1 through 5, wherein the model storage unit further stores a value indicating demand for each of a plurality of three-dimensional homologous models classified into a plurality of clusters, based on the number of the three-dimensional homologous models in each cluster for a plurality of clusters classified by cluster analysis.
A body shape analysis method, including:
| Number | Date | Country | Kind |
|---|---|---|---|
| 2023073079 | Apr 2023 | JP | national |
