JP2018164727A - Information processing device, information processing method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To equally and easily enable matching of even items having unstandardized shapes.SOLUTION: An information processing device includes measurement means for specifying a correct position on the basis of measurement data of an item, detecting one or more landmark points, calculating a reference symbol on the basis of a prescribed rule for removing designability of the item, calculating each measurement value of a plurality of items on the basis of the correct position, the landmark points, and the reference symbol, and generating information including at least each measurement value of the plurality of items as measurement data of the item.SELECTED DRAWING: Figure 5

Description

  The present invention relates to an information processing apparatus, an information processing method, and a program.

  Conventionally, the standardization of shoes has been delayed because techniques such as making shoes according to wooden shapes handmade by craftsmen have been taken. In order to perform optimal matching, it is necessary to know the inner dimensions of the shoe in consideration of the elongation and deformation of the shoe, but the conventional method cannot quantitatively determine the degree of elongation or deformation of the shoe. Further, since the error is large, it is necessary to use a destructive method, which is costly.

JP 2012-033128 A

  For this reason, there is a need for a new technique that can be used to measure items equally and easily for items such as shoes whose shapes are not standardized.

  The present invention has been made in view of such a situation, and an object of the present invention is to provide a new technique capable of performing matching evenly and easily even with an item whose shape is not standardized. .

In order to achieve the above object, an information processing apparatus of one embodiment of the present invention provides:
Based on the measurement data of the item, identify the correct pair, detect one or more landmark points, calculate a reference symbol based on a predetermined rule for eliminating the design of the item, Measuring means for calculating each measurement value of a plurality of items based on a landmark point and the reference symbol, and generating information including at least each measurement value of the plurality of items as measurement data of the item .

  According to the present invention, even an item whose shape is not standardized can be matched equally and easily.

It is a block diagram which shows the structural example of the information processing system containing the information processing apparatus with which this invention is applied. It is the conceptual diagram which showed typically about processes, such as machine learning used as the premise of the matching in this service to which the information processing system shown in FIG. 1 is applied. 1 is a configuration diagram of an information processing system including a matching device that is an embodiment of an information processing device of the present invention. It is the conceptual diagram which showed typically about processes, such as matching in this service to which the information processing system shown in FIG. 1 is applied. It is a block diagram which shows an example of the hardware constitutions of the matching apparatus 5 among the information processing systems of FIG. 1 which can implement | achieve this service (matching). 1 is a functional block diagram showing an example of a functional configuration capable of realizing the processing such as machine learning shown in FIG. 2 among the processing executed by the information processing system 1 shown in FIG. 1 including the matching device 5 having the hardware configuration shown in FIG. It is. It is a figure which shows the outline of the external appearance structure of the leg for demonstrating the standardization foot measurement method. It is a figure which shows the outline of the external appearance structure of the shoe and shoe anchor for demonstrating the standardized shoe measurement method. It is a functional block diagram which shows an example of a functional structure which can implement | achieve processes, such as matching shown in FIG. 3, among the processes which the information processing system 1 of FIG. 1 performs. It is a figure explaining an example of the matching method determined based on the result of machine learning. FIG. 4 is a diagram illustrating an outline of processing that can be performed by the information processing system of FIG. FIG. 3 is a diagram illustrating an outline of processing that can be performed by the information processing system of FIG. 1 and performs machine learning different from the example of FIG. 2. An example of measurement data of a user U (including a tester T) when standardized measurement is performed when the item is clothing is shown. An example of measurement data of an item when standardized measurement is performed when the item is clothes (jacket) is shown. It is an example of the mock of the screen of EC site. It is an example of the mock of the screen of EC site. It is an example of the mock of the screen of EC site.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram illustrating a configuration example of an information processing system including an information processing apparatus to which the present invention is applied.
The information processing system shown in FIG. 1 includes a shoe measurement device 1, a tester foot measurement device 2, a learning device 3, a user foot measurement device 4, a matching device 5, a user terminal 6, a shoe measurement DB 11, and a foot It is comprised so that measurement DB12 and evaluation DB13 may be included.
The shoe measuring device 1 to the user terminal 6 and the shoe measuring DB 11 to the evaluation DB 13 are connected to each other via a predetermined network N such as the Internet.

The outline of a service (hereinafter referred to as “the present service”) realized using the information processing system of FIG. 1 will be described below.
This service is for measuring items that are not standardized (for example, shoes) evenly and easily, and matching with users who wish to purchase (recommendation of recommended shoes, etc.) It is a service that can. According to this service, as will be described later, since the trend seen in each of the user's body shape and item shape is captured abstractly, matching is done easily and equally without necessarily comparing the same position. It can be performed. Furthermore, since errors at the time of measurement are statistically captured, practical matching can be performed even for measurement values obtained by inexpensive sensors that are likely to cause errors.
The item to which this service is applied is not particularly limited as long as it is attached to at least a part of the user's body, but in the following example, it is assumed to be shoes.

  FIG. 2 is a conceptual diagram schematically showing processing such as machine learning, which is a premise of matching in this service to which the information processing system shown in FIG. 1 is applied.

Here, when modeling a matching target, measurement may be performed using actual shoes, but in the present embodiment, m (m is an integer value of 1 or more) shoe anchors Sh1 to Shm are used. When it is not necessary to individually distinguish each of the shoe anchors Sh1 to Shm, these are collectively referred to as “shoe anchor Sh” hereinafter.
In step Sa, the shoe measuring device 1 generates each of the 3D scan data Ds1 to Dsm by scanning each of the shoe anchors Sh1 to Shm.
In step Sb, the shoe measuring device 1 measures the shoe (as will be described in detail later) for each of the 3D scan data Ds1 to Dsm, which is devised by the present inventors and adopted as a standard one in this embodiment. However, hereinafter, measurement is performed according to “standardized shoe measurement method”), and data obtained as a result of the measurement (hereinafter referred to as “shoe measurement data”) Ks1 to Ksm is stored in the shoe measurement DB 11 Let
Hereinafter, when it is not necessary to individually distinguish each of the shoe measurement data Ks1 to ksm, these are collectively referred to as “shoe measurement data Ks”.

On the other hand, n people (n is an integer value of 1 or more) natural people T1 to Tn are asked to actually wear m shoes for evaluation. Hereinafter, each of such natural persons T1 to Tn is hereinafter referred to as “testers T1 to Tn”. When it is not necessary to individually distinguish each of the testers T1 to Tn, these are collectively referred to as “tester T” hereinafter.
In step Sc, the tester foot measuring apparatus 2 generates 3D scan data Dt1 to Dtn by scanning both feet of the testers T1 to Tn.
In step Sd, the tester foot measuring apparatus 2 measures the foot (as to details) which is devised by the present inventors and adopted as a standard one in this embodiment for each of the 3D scan data Dt1 to Dtn. , Which will be described later, hereinafter referred to as “standardized foot measurement method”), and data obtained as a result of the measurement (hereinafter referred to as “foot measurement data”) Kt1 to Ktm are each measured in the foot measurement DB12. Remember me.
Hereinafter, when there is no need to individually distinguish each of the foot measurement data Kt1 to ktn, these are collectively referred to as “foot measurement data Kt”.
Further, for each of the testers T1 to Tn, evaluation data (hereinafter simply referred to as “evaluation data”) Ht1 to Htn relating to the comfort when the m shoes are actually worn are acquired and stored in the evaluation DB 13. Stored.
Hereinafter, when it is not necessary to individually distinguish each of the evaluation data Ht1 to Htn, these are collectively referred to as “evaluation data Ht”.

The learning device 3 includes each combination of shoe measurement data Ks1 to Ksm of m shoes (shoe anchors Sh1 to Shm) and each of foot measurement data Kt1 to Ktn of n testers T1 to Tn, Machine learning is performed using the evaluation data Ht1 to Htn for each of the m shoes of each of the n testers T1 to Tn.
Thereby, as a learning result, for example, a tendency of comfort is obtained for each of various combinations of a predetermined foot and a predetermined shoe. It should be noted that the shoe measurement data Ks and the foot measurement data Kt do not necessarily need to include data measured at the same position. In other words, even if measurement is performed at different positions, a machine-learning can provide a tendency of comfort for various combinations of a predetermined foot and a predetermined shoe.
By using such learning results, it is possible to grasp the trend of the shape of each foot and shoe in an abstract manner, and matching (details will be described later) without necessarily comparing the same position. Is possible.
In addition, by using such a machine learning method, it is possible to statistically handle errors during measurement, so that practical matching is possible even in scans using inexpensive sensors with large errors.

  FIG. 3 is a conceptual diagram schematically showing processing such as matching in this service to which the information processing system shown in FIG. 1 is applied.

In step Sf, the user foot measurement device 4 generates 3D scan data Du by scanning both feet of the user U.
In step Sg, the user foot measurement device 4 performs measurement according to the standardized foot measurement method for the 3D scan data Du, and stores the foot measurement data Ku obtained as a result of the measurement in the foot measurement DB 12.

In step Si, the matching device 5 uses the foot measurement data Ku of the user U obtained in step Sg, and the shoe measurement data Ks1 to Ks1 of the m shoes (shoe anchors Sh1 to Shm) described above with reference to FIG. Using each of Ksm, matching is performed for estimating which shoes are comfortable for the user U.
Here, not only the comparison of measurement data, but also the matching device 5 may acquire the preference of the user U in Step Sh and perform matching in consideration of the preference in Step Si. Thereby, for example, when tight shoes are the preference of the user U, even if it is generally determined that the A size shoes are comfortable to wear, a B size smaller than the A size is suitable for the user U. Matching suitable for each user U, such as being determined to be present, becomes possible.
In step Sj, the matching device 5 sends, to the user U (user terminal 6), shoes suitable for the user U (the user U will feel comfortable) based on the matching result of step Si. Recommended for.

If necessary, the comfort of the user U when actually wearing shoes recommended by the user U (shoe anchor Sh: h is an arbitrary integer value from 1 to m) is evaluated. In this case, the evaluation data Hu of the user U is stored in the evaluation DB 13.
When the evaluation data Hu of the user U, the foot measurement data Ku of the user U, and the shoe measurement data Ksk of the recommended shoe (shoe anchor Sh) are provided to the learning device 3, the learning device 3 , Perform machine learning (relearning). As described above, the actual recommendation result and its evaluation are fed back and re-learning is performed, so that matching with higher accuracy becomes possible.

  FIG. 4 is a block diagram showing an example of a hardware configuration of the matching device 5 in the information processing system of FIG. 1 capable of realizing such a service (matching).

  The matching device 5 includes a CPU (Central Processing Unit) 21, a ROM (Read Only Memory) 22, a RAM (Random Access Memory) 23, a bus 24, an input / output interface 25, an output unit 26, and an input unit 27. A storage unit 28, a communication unit 29, and a drive 30.

The CPU 21 executes various processes according to a program recorded in the ROM 22 or a program loaded from the storage unit 28 to the RAM 23.
The RAM 23 appropriately stores data necessary for the CPU 21 to execute various processes.

  The CPU 21, ROM 22, and RAM 23 are connected to each other via a bus 24. An input / output interface 25 is also connected to the bus 24. An output unit 26, an input unit 27, a storage unit 28, a communication unit 29, and a drive 30 are connected to the input / output interface 25.

The output unit 26 includes a display, a speaker, and the like, and outputs various types of information as images and sounds.
The input unit 27 includes a keyboard, a mouse, and the like, and inputs various information.

The storage unit 28 includes a hard disk, a DRAM (Dynamic Random Access Memory), and the like, and stores various data.
The communication unit 29 communicates with other devices via the network N including the Internet.

A removable medium 31 composed of a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is appropriately attached to the drive 30. The program read from the removable medium 31 by the drive 30 is installed in the storage unit 28 as necessary.
The removable medium 31 can also store various data stored in the storage unit 28 in the same manner as the storage unit 28.

  In addition, each of the shoe measuring device 1, the tester foot measuring device 2, the learning device 3, the user foot measuring device 4, and the user terminal 6 can adopt the same hardware configuration as FIG. Therefore, description of these hardware configurations is omitted.

  FIG. 5 shows an example of a functional configuration capable of realizing the processing such as machine learning shown in FIG. 2 among the processing executed by the information processing system 1 of FIG. 1 including the matching device 5 having such a hardware configuration. It is a functional block diagram shown.

As shown in FIG. 5, in the shoe measuring device 1, a scanning unit 111 and a standardized shoe measuring unit 112 function.
In the tester foot measurement apparatus 2, the scanning unit 121, the standardized foot measurement unit 122, and the evaluation acquisition unit 123 function.

The scanning unit 111 of the shoe measuring device 1 generates 3D scan data Ds by scanning the shoe anchor Sh.
The standardized shoe measurement unit 112 performs measurement according to the standardized shoe measurement method for the 3D scan data Ds, and stores the shoe measurement data Ks obtained as a result of the measurement in the shoe measurement DB 11.
Here, the shoe measurement data Ks of the shoe (or shoe anchor Sh) can be stored in the shoe measurement DB 11 by associating various information related to the shoe. Specifically, for example, 3D scan data Ds can be associated with shoe measurement data Ks. Further, for example, information on the type of material for each part of the shoe and information on the properties of the material such as tensile strength (specific strength) can be linked to the shoe measurement data Ks.

On the other hand, the scanning unit 121 of the tester foot measurement device 2 generates 3D scan data Dt by scanning both feet of the tester T.
The standardized foot measurement unit 122 performs measurement according to the standardized foot measurement method for the 3D scan data Dt, and stores the foot measurement data Kt obtained as a result of the measurement in the foot measurement DB 12.
Here, various information regarding the tester T can be associated with the foot measurement data Kt of the tester T. Specifically, for example, the 3D scan data Dt can be associated with the foot measurement data Kt. Also, for example, gender, age, weight, average number of steps per day, shoes purchase history and return history, various types of shoes such as frequently worn shoes, uninterested shoes, favorite brands, disliked brands, etc. Such information can be linked to the foot measurement data Kt. Further, for example, information such as the address and occupation of the tester T may be associated with the foot measurement data Kt. In addition, when many foot measurement data Kt and the above-mentioned shoe measurement data Ks are obtained, these data can be utilized as various kinds of applications such as various analyzes as big data.
In addition, the evaluation acquisition unit 123 acquires evaluation data Ht indicating evaluation regarding the comfort when the tester T actually wears each of the shoes, and stores the evaluation data Ht in the evaluation DB 13.

Here, the standardized foot measurement method will be described with reference to FIG. 6, and the standardized shoe measurement method will be described with reference to FIG.
FIG. 6 is a diagram showing an outline of the external appearance configuration of a foot for explaining the standardized foot measurement method.
FIG. 7 is a diagram showing an outline of the external configuration of shoes and shoe anchors for explaining a standardized shoe measuring method.

  The standardized foot measurement unit 122 of the tester foot measurement device 2 performs measurement according to the standardized foot measurement method, for example, as follows.

The standardized foot measurement unit 122 identifies “directly facing”.
That is, the standardized foot measurement unit 122 uses PCA (principal component analysis) for the point group of the 3D scan data Dt of the tester T's foot to obtain a direction in which the distribution of the data is large, and determines the facing direction Determined uniquely. This eliminates the need for the tester T to precisely align the feet during measurement.
In other words, in order to extract and compare lengths and angles from a foot that can take various shapes according to a certain rule, it is necessary to first determine a reference axial direction. Therefore, the standardized foot measurement unit 122 defines the axis in which the foot footprint shape is most widely distributed in the 3D scan data Dt as the traveling direction axis, and takes the gravity direction as the vertical axis as it is, By defining the orthogonal axis as the left and right axis, the “face-to-face” is specified.

The standardized foot measurement unit 122 detects the “landmark point L”.
That is, the standardized foot measurement unit 122 detects landmark points L such as toes and heel vertices using the characteristics of the shape of the foot. For points that are difficult to extract only by shape characteristics (ball joint (toe axis), base of thumb, end point of thumb and little finger), a colored marker is pasted on the tester T's foot during measurement, and a standardized foot measurement unit 122 is determined based on this.

The standardized foot measurement unit 122 obtains a “measurement value”.
That is, the standardized foot measurement unit 122 performs geometric calculation based on the landmark point L thus obtained, and handles the value obtained by the calculation as a standardized measurement value.

The measured values are as follows.
“Foot circumference A”: circumference of the foot when the foot is sliced with the ball joint J (L).
“Foot width B”: A linear distance between the ball joints J (L).
“Foot length C”: The length of the toe (frontmost point) and the heel apex (most rearmost point) on the floor plane.
“Wheel shape D”, “Toe height E”, “Toe shape F”, and “Distortion G”.
In addition, the following measured values may be used as necessary.
“Long length”: Length at a certain ratio to the foot length C.
“Head width”: The width of the foot on the surface that extends forward from the last point by the length of the heel.
“Deep Depth”: Distance (depth) from the heel apex position to the foot, extending vertically upward by a specified length.
“Toe Width”: Distance between the end of the thumb base and the end of the little finger base.
“Height of thumb base, thumb, little finger”: Height from each landmark point L to the floor.

On the other hand, the standardized shoe measuring unit 112 of the shoe measuring device 1 performs measurement according to the standardized shoe measuring method as follows, for example.
That is, there are various conventional methods for measuring the inner dimensions of shoes (photographing by CT, mold making using plaster, measurement of a mold (wood pattern) at the time of manufacture, etc.). However, when machine learning is used as in the present embodiment, it is only necessary to know the relative differences that characterize the individual, so it is sufficient to obtain the following “measurement values”. In other words, any method (various name methods) for obtaining the following “measurement values” is the standardized shoe measurement method.

  In obtaining the “measurement value”, the standardized shoe measurement unit 112 identifies “facing” and detects the “landmark point L” in the same manner as the standardized foot measurement method described above with reference to FIG.

The standardized shoe measuring unit 112 also obtains the next reference symbol (line, surface).
The following reference symbols are required according to the following rules in order to eliminate the design of shoes.
"Toe position"
(Rule) Since there is a hollow portion where a foot cannot enter due to the design, a position where the foot circumference A is constant (for example, 100 mm) is defined as a toe position in order to eliminate the influence of that portion.
"Ball joint (joint part of bone connecting toe and heel)"
(Rule) The point farthest in the horizontal direction corresponds to the ball joint of the foot.

  The standardized shoe measuring unit 112 performs “foot circumference A”, “foot width B”, “foot length C”, “foot length C” based on such “facing”, “landmark point L”, and “reference symbol”. The heel shape D, “toe height E”, “toe shape F”, “distortion G”, and the like are obtained as “measurement values”.

According to the standardized shoe measurement method, it is necessary to determine the reference axial direction (facing) for shoes in the same way as the standardized foot measurement method. Further, unlike a foot having a bone inside, since the position of the ball joint is indefinite, it is necessary to be able to acquire the position stably.
Therefore, as shown in FIG. 7, a shoe anchor Sh having a shape that exactly fits the toe of the foot is created, and three-dimensional measurement (scanning) is performed together with the shoe anchor Sh packed in the shoe. Thereby, a reference axis (facing) and a reference symbol based on the shoe anchor Sh are obtained, and based on these, each measurement value is obtained appropriately.
As described above, the standardized shoe measurement unit 112 can construct a model (shoe measurement data Ks) from which design elements such as toes are excluded by using the shoe anchor Sh as a reference.

  As the shoe anchor Sh, an arbitrary one can be adopted as long as the above-mentioned reference axis (facing) and reference symbol can be obtained, but a shape as shown in FIG. 7 is preferable. That is, as shown in FIG. 7, it is not necessary to prepare the shoe anchor Sh for each height of the heel by creating up to the ball joint portion. The toe shape of the shoe anchor Sh adopts an average shape in the example of FIG. 7, but it is more preferable to make a variation using a statistical model.

As described above, the standardized shoe measurement method and the standardized foot measurement method are methods for obtaining a plurality of “measurement values” newly proposed by the present inventors. Machine learning is performed on the basis of the plurality of “measurement values”, and the matching between the foot (of the new user U) and the shoes using the learning result is performed.
That is, the standardized shoe measurement method and the standardized foot measurement method are methods obtained by standardizing the measurement length that affects matching from the shape of shoes and feet that have not been standardized. That is, a standardized shoe measurement method or a standardized foot measurement method is a measurement method that can be compared equally regardless of what type of shoe or foot is extracted.
Therefore, since machine learning is performed using the shoe measurement data Ks obtained by the measurement of the standardized shoe measurement method and the foot measurement data Kt obtained by the measurement of the standardized foot measurement method, each measurement value is obtained as the learning result. The relationship (weighting) that has been determined by the craftsman in the past is automatically obtained as to which value has an effect on matching (such as comfort).

Here, the difference between the measurement value obtained by the standardized shoe measurement method (abbreviated as “shoes” in this paragraph) and the measurement value obtained by the standardized foot measurement method (abbreviated as “foot” in this paragraph) Will be described.
Regarding “foot circumference A”, “foot” has a contour line length or shape obtained by cutting the ball joint portion perpendicularly to the ground, whereas “shoes” is the contour of the corresponding portion of shoe anchor Sh that fits perfectly. Line length and shape.
Regarding “foot width B”, “foot” is the line length connecting the ball joint part horizontally to the ground, whereas “shoes” is the horizontal line length of the most protruding part.
Regarding “foot length C”, “foot” is the length of the toe at the tip of the toe and the rear end of the heel with respect to the traveling direction axis, whereas “shoe” is (shoe anchor Sh From the reference symbol to the toe of the shoe anchor Sh) + (path length along the insole from the reference symbol of the shoe anchor Sh to the heel end).
As for “heel shape D”, “foot” is a change in the contour shape when it is cut horizontally on the ground while shifting in the height direction with reference to the heel end, whereas “shoes” (Change in contour shape when horizontally cut on the ground while shifting in the height direction with respect to the rear end of the heel) + (shape of the mouth portion).
Regarding “toe height E”, “foot” is the height (average or distribution) of the foot at a certain length before the line of the ball joint, whereas “shoe” is the ball joint. Is the average height of the foot in a portion that is a certain length before the line (the bottom surface is the insole top surface).
As for “toe shape F”, “foot” has a trapezoidal shape (angle, length, etc.) connecting a ball joint, thumb, and little finger, whereas “shoes” have a certain length above the upper surface of the insole. It is a contour shape when the toe portion is cut into a ring.
Regarding “distortion G”, “foot” is a relative angle from the ball joint to the tip of the heel, whereas “shoe” is a relative angle from the ball joint to the tip of the heel.

As described above, the shoe measurement data Ks1 to Ksm of m shoes (the shoe anchors Sh1 to Shm) are obtained by the standardized shoe measurement method. Further, the foot measurement data Kt1 to Ktn of n testers T1 to Tn are obtained by the standardized foot measurement method.
Then, the learning device 3 in FIG. 5 performs each of the shoe measurement data Ks1 to Ksm of m shoes (shoe anchors Sh1 to Shm) and the foot measurement data Kt1 to Ktn of n testers T1 to Tn. Machine learning is performed using each of the combinations and the evaluation data Ht1 to Htn for each of the m shoes of the n testers T1 to Tn.

Next, a functional configuration capable of realizing the processing such as matching shown in FIG. 3 performed using the learning result of the learning device 3 among the processing executed by the information processing system of FIG. 1 will be described.
FIG. 8 is a functional block diagram showing an example of a functional configuration capable of realizing the processing such as matching shown in FIG. 3 among the processing executed by the information processing system 1 of FIG.

  As shown in FIG. 8, in the user foot measurement device 4, a scan unit 141, a standardized foot measurement unit 142, and an evaluation acquisition unit 143 function.

The scan unit 141 of the user foot measurement device 4 generates 3D scan data Du by scanning both feet of the user U.
The standardized foot measurement unit 122 performs measurement according to the above-described standardized foot measurement method for the 3D scan data Du, and stores the foot measurement data Ku obtained as a result of the measurement in the foot measurement DB 12.
Here, as with the tester T, various information related to the user U can be associated with the foot measurement data Ku of the user U. Specifically, for example, the 3D scan data Du can be associated with the foot measurement data Ku. Also, for example, gender, age, weight, average number of steps per day, shoes purchase history and return history, various types of shoes such as frequently worn shoes, uninterested shoes, favorite brands, disliked brands, etc. Such information can be linked to the foot measurement data Ku. Further, for example, information such as the address and occupation of the user U may be associated with the foot measurement data Ku. The foot measurement data Ku of the user U can be used together with the measurement data Kt of the tester T for various purposes such as various analyzes as a part of big data.
In addition, the evaluation acquisition unit 143 acquires evaluation data Hu indicating evaluation about the comfort when the user U actually wears the shoes recommended as a result of matching by the matching device 5 from the user terminal 6 or the like, and evaluates it. Store in DB13.

  For such a user foot measurement device 4, in the matching device 5, a matching method determination unit 151, a user preference acquisition unit 152, a matching unit 153, and a recommendation unit 154 function.

The matching technique determination unit 151 determines an appropriate matching technique for the user U based on the result of machine learning by the learning device 3.
The user preference acquisition unit 152 acquires the user U's preference regarding shoes from the user terminal 6 or the like.
The matching unit 153 uses the user U's foot measurement data Ku, each of the shoe measurement data Ks1 to Ksm of each of the m shoes (shoe anchors Sh1 to Shm), and the user U's preference to use the user U's preference. Matching (to estimate) which shoes are comfortable to wear is performed.
Based on the matching result of the matching unit 153, the recommendation unit 154 recommends shoes suitable for the user U (the user U will feel comfortable to wear) to the user U (user terminal 6). To do.

Here, with reference to FIG. 9, an example of the matching technique determined by the matching technique determination unit 151 based on the result of machine learning by the learning device 3 will be described.
FIG. 9 is a diagram for explaining an example of a matching method determined based on the result of machine learning.
As shown in FIG. 9, for each of the testers T1 to Tm, a combination of the shoe measurement data Ksk of the predetermined shoe (shoe anchor Sh) and the foot measurement data Kt of the tester T, and the predetermined shoe is actually worn. Machine learning by the learning device 3 is performed using a set with the evaluation data Ht at that time.
Here, as described above, each of the shoe measurement data Ksk and the foot measurement data Kt of the tester T is a set of a plurality of measurement values such as “foot circumference A”, “foot width B”, and “foot length C”. Is the body. The shoe evaluation data Ht is also a collection of individual evaluations of the comfort for each item corresponding to a plurality of measurement values such as “foot circumference A”, “foot width B”, and “foot length C”. .
Therefore, when machine learning as shown in FIG. 9 is performed, the prediction of the evaluation of the user U with respect to the combination of the shoe measurement data Ksk of the predetermined shoe (the shoe anchor Sh) and the foot measurement data Ku of the user U is predicted. A matching method for performing is used. Here, the prediction of the user's evaluation is a prediction of the individual evaluation of the comfort for each item corresponding to a plurality of measurement values such as “foot circumference A”, “foot width B”, and “foot length C”.

  Although one embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment, and modifications, improvements, and the like within the scope that can achieve the object of the present invention are included in the present invention. It is.

  In the above-described embodiment, machine learning is performed on the relationship between the foot of the tester U and the m shoes, and the foot of the user U and the shoe are matched based on the result of the machine learning. Based on the result of the matching Thus, shoes suitable for the feet of the user U (shoes that are estimated to be highly evaluated by the user U) were recommended, but the target of machine learning (matching and recommendation using the result of the machine learning is particularly important) It is not limited.

For example, the machine learning target (matching and recommendation target using the machine learning result) may be as shown in FIG.
FIG. 10 is a diagram showing an outline of processing that can be performed by the information processing system of FIG. 1 and is different from the example of FIG.
As shown in FIG. 10, for each of m existing shoes S1 to Sm (shoe anchors Sh1 to Shm), shoe measurement data Ks1 to Ksm and a large number of users U1 to Uk (k is an arbitrary integer value, FIG. In this example, shoes are matched based on a combination with evaluation data Hu (including evaluation data Ht of the tester T) of k = 3), and existing shoes with similar evaluation to new shoes are recommended. .
That is, each of m existing shoes (the shoe anchors Sh1 to Shm, which are the matching comparison destinations) on a two-dimensional plane with the shoe measurement data Ks as the A axis and the evaluation data Hu as the B axis. Are plotted (projected) based on the combination of the shoe measurement data Ks1 to Ksm and a large number of users U1 to Uk, and machine learning is performed. In other words, machine learning is performed so that a two-dimensional plane (space) in which shoes (items) having similar evaluations are arranged nearby is generated.
As a result of this machine learning, a new shoe (matching comparison source) is plotted (projected) on the two-dimensional plane (space) based on a combination of shoe measurement data Ks and a large number of users U1 to Uk. Thus, it is possible to search (match) shoes arranged near the two-dimensional plane (space) and recommend them as shoes having similar evaluations, that is, shoes having similar comfort.

Further, for example, the machine learning target (matching and recommendation target using the machine learning result) may be the one shown in FIG.
FIG. 11 is a diagram showing an outline of processing that can be performed by the information processing system of FIG. 1 and performs machine learning different from the example of FIG. 2.
In the machine learning shown in FIG. 3, a model (existing model) based on a general evaluation of n (many) testers T for m shoes is obtained.
On the other hand, in the example of FIG. 11, feedback information from the user U that has been matched and recommended using these existing models, for example, information indicating various histories (purchasing history and return history) regarding the user U is also considered. Then, additional learning is performed. As a result of this additional learning, a model specialized to the preference of the user U (compared to the existing model), that is, a user-specific model is obtained.
By performing matching using this user-specific model, it is possible to recommend shoes that are more suitable for the user U than when using an existing model.

  In addition, for example, the machine learning target (matching and recommendation target using the machine learning result) is a shoe in the above-described embodiment, but is not particularly limited to this, and is worn by the user U (including the tester T). It may be any possible item, such as clothes, hats, helmets for protection or assistance, corsets, supporters and the like.

FIG. 12 shows an example of measurement data of the user U (including the tester T) when standardized measurement is performed when the item is clothing.
That is, since the measurement method of the human body (the body of the user U) is standardized, the measurement by the measurement method is adopted as the standardized measurement, and the measurement result is adopted as the standardized measurement value, that is, the user U An example of the measurement data is shown in FIG.

FIG. 13 shows an example of item measurement data when standardized measurement is performed when the item is clothing (jacket).
That is, since the measurement method such as which part of the clothing (jacket) is measured for each category is standardized, measurement by the measurement method is adopted as standardized measurement, and the measurement result is standardized measurement value. FIG. 13 shows an example of adopted data, that is, an example of measurement data of clothes (jacket).

In addition, for example, in the above-described embodiment, matching based on the shape of the item is performed. However, the present invention is not particularly limited thereto, and matching in consideration of other elements may be performed together with the shape or separately from the shape. . In other words, the factors that determine the comfort of an item (the comfort of an item if it is a shoe) are not only the difference in shape, but if it is a shoe, there are factors such as weight, age, heel height, and shoe material. In the case of clothes, there are factors such as the material of the fabric, sex, and age.
In the past, when considering the influence of these “other elements”, it was necessary to attach rules to all the influences. On the other hand, by adopting a machine learning method (based on a decision tree) that can automatically learn the influence of multiple elements, the influence of "other elements" can be achieved without a rule. Matching that takes into account can be realized.
Here, as the “other elements”, the following can be adopted.
In other words, it is described as follows in the format of “other elements” = (data type, influence).
“Ease of material extension” = (Tensile strength [N / cm], affecting the tolerance of shape difference)
“Material type” = (Category (cloth, polyester, etc.), affects the pain of the contact area)
"Body weight" = (kg, heel affects the load on the toes)
"Heel height" = (cm, affects the load on the toes at the heel)
"Age" = (Influence on preference, such as numerical value and ease of walking)
Affects differences in preferences such as daily average number of steps and fatigue

In addition, for example, in the above-described embodiment, indirect machine learning using an evaluation when the tester T actually wears m shoes is performed, but the present invention is not particularly limited to this, and together with or instead of this evaluation. For example, using a cloth-type pressure distribution sensor, the pressure applied when each of the m shoes is worn by the tester T is measured, and machine learning is performed using the measurement results. The prediction may be made even when there is no pressure distribution sensor.
Also in this case, the estimation of the final preference of the user U (the comfort in the case of shoes) is performed using the prediction result, and the recommendation to the user U based on the estimation result is performed.

Further, for example, in the above-described embodiment, various pieces of information (metadata) are associated with the shoe measurement data Ks of the shoe, the foot measurement data Kt of the tester T, and the foot measurement data Ku of the user U. This is because machine learning, matching (preference estimation), and recommendation are performed with higher accuracy by considering these various pieces of information (metadata).
Such metadata is not limited to the above-described embodiment, and any metadata can be adopted. For example, the following metadata can be adopted. Note that the metadata that is marked with (better) in parentheses is metadata that is preferably adopted. Although it is not essential to adopt metadata that is marked with (advanced) in parentheses, it is possible to perform more appropriate machine learning, matching (preference estimation), and recommendation if adopted. .
That is, as the material property metadata, elongation (better), surface hardness (better), and material type (enamel, smooth, cloth, etc.) (advance) can be adopted.
Brand name, manufacturer name, factory name (advance), and manufacturing method (advance) can be adopted as metadata of manufacturing information.
As the attribute information of the user U, a proficiency level (better), a usage frequency (better), and an age (advance) can be adopted.

In the above-described embodiment, the user U visits a shoe store, where his / her foot is scanned, and matching is performed using the foot measurement data Ku obtained as a result. Based on the result, the user U Appropriate shoes for U (shoes predicted to be comfortable) were recommended.
However, it is not particularly essential to use the foot measurement data Ku when matching.
For example, in order to recommend shoes and clothes in the state where there is no user U's foot / body shape data on an EC site or the like, for example, matching is performed using shoes or clothes that the user U has purchased in the past, Those having the same or similar shape may be searched and recommended.
In this case, there are a large number of types of measurement values of shoes and clothes, and there is a possibility that numerical values important for fitting are not always emphasized by their simple similarity. Therefore, the information processing system to which the above machine learning is applied is adopted, the degree of contribution to the similarity to the measured values of shoes and clothes is calculated using the similarity of the prior evaluation by the tester T (test user), It is preferable to perform a search that reflects the degree of contribution.

Specifically, for example, evaluation by a combination of a tester T1 and shoes (hereinafter referred to as “first combination”) is (foot length difference, knob thickness difference, heel width difference, ease of removal) = (+ 0.1, -0.8, +0.1, A). Further, the evaluation by another tester T2 and shoe combination (hereinafter referred to as “second combination”) is (foot length difference, knob thickness difference, heel width difference, ease of removal) = (+ 0.1, −0 .8, +0.1, A).
In this case, in order to estimate ease of removal by machine learning, it is assumed that the ratio of important items (contribution) is calculated as follows, for example. That is, it is assumed that contribution = (foot length difference, knob thickness difference, heel width difference) = (0.3, 0.1, 0.6).
Using this contribution degree, weighting of the similarity index of shoes (each measured value of the shoe measurement data Ks) can be weighted to perform matching (search).
For example, the shoe measurement data Ks1 of the first shoe is (foot length, knob thickness, heel width) = (23.2, 2.4, 4.55), and the shoe measurement data Ks1 of the second shoe is ( It is assumed that foot length, knob thickness, heel width) = (23.6, 3.6, 4.6). In this case, in the above example, since the contribution degree of the heel width difference is the highest at 0.6, matching (search) is performed by placing importance on the heel width of the measured values. That is, the heel width is 4.55 for the first shoe and 4.6 for the second shoe, and therefore, the first shoe and the second shoe are determined to be very similar in terms of ease of removal. In other words, if the user U purchases the first shoe in the past, the second shoe is recommended for the user U who places importance on ease of removal.

The screen of the EC site displayed on the user terminal 6 when such a search (recommendation) is performed, and the foot of the user U is scanned at the store and displayed on the store side device (for example, the user foot measuring device 4). The screens to be displayed are as shown in FIGS. 14 to 16, for example.
FIG. 14 is a diagram illustrating an example of a mock of a recommendation screen at a store.
FIG. 15 is an example of a mock of a recommendation screen at a store, and shows an example of a screen displayed after the display of the screen of FIG.
FIG. 16 is an example of a mock of the screen of the EC site.

  For example, each hardware configuration shown in FIG. 4 is merely an example for achieving the object of the present invention, and is not particularly limited.

  Further, the functional block diagrams shown in FIGS. 5 and 8 are merely examples, and are not particularly limited. That is, it is sufficient that the information processing system has a function capable of executing the above-described series of processing as a whole, and what functional blocks are used to realize this function are shown in the examples of FIGS. There is no particular limitation.

Further, the location where the functional block exists is not limited to the location shown in FIG. 5 or 8 and may be arbitrary. For example, at least a part of the functional blocks of each device may be provided on the user terminal 6 side. That is, as viewed from the user terminal 6, an information processing system that causes an external server or the like to execute processing using a browser may be constructed, or an information processing system that causes an installed application program to execute processing may be constructed. Alternatively, an information processing system composed of these combinations may be constructed.
One functional block may be constituted by hardware alone or in combination with software alone.

When the processing of each functional block is executed by software, a program constituting the software is installed on a computer or the like from a network or a recording medium.
The computer may be a computer incorporated in dedicated hardware. The computer may be a computer capable of executing various functions by installing various programs, for example, a general-purpose smartphone or personal computer other than a server.

  The recording medium including such a program is not only constituted by a removable medium distributed separately from the apparatus main body in order to provide the program, but is provided to each user in a state of being preinstalled in the apparatus main body. It consists of a recording medium.

  In the present specification, the steps for describing the program recorded on the recording medium are not limited to the processing performed in time series according to the order, but may be performed in parallel or individually even if not necessarily performed in time series. The process to be executed is also included.

In summary, the information processing apparatus to which the present invention is applied only needs to have the following configuration, and can take various embodiments.
That is, the information processing apparatus to which the present invention is applied (for example, the matching apparatus 5 in FIG. 1)
First type standard data (for example, shoe measurement data Ks1 in FIG. 2) obtained by standardizing measurement data (for example, 3D scan data Ds1 to Dsm in FIG. 2) of a predetermined item (for example, shoes, which may be shoe anchors Sh1 to Shm). To Ksm) and second type standard data (for example, 3D scan data Dt1 to Dtn in FIG. 2) standardized from measurement data (for example, 3D scan data Dt1 to Dtn in FIG. 2) of a predetermined tester (for example, testers T1 to Tn in FIG. 2). Based on the result of machine learning (for example, step Se in FIG. 2) using the combination with the foot measurement data Kt1 to Ktn in FIG. 2 for each of the plurality of items and the plurality of testers, the user (for example, the user U in FIG. 3). 8) predicting means for predicting the degree of suitability of the item (for example, matching in FIG. 8 for matching in step Si in FIG. 3) And part 153),
Recommending means for recommending an item to the user based on the prediction result of the predicting means (for example, the recommendation unit 154 in FIG. 8 that recommends in step Sj in FIG. 3);
Is provided.

In addition, an information processing apparatus according to another aspect to which the present invention is applied (for example, the shoe measurement apparatus 1 in FIG. 1)
Based on the measurement data of the item (for example, the 3D scan data Ds1 to Dsm of FIG. 2 generated by the scan unit 111 of FIG. 5), the correctness is identified, one or more landmark points are detected, and the item design A reference symbol is calculated based on a predetermined rule for eliminating the sex, and each measurement value of a plurality of items is calculated based on the front, the landmark point, and the reference symbol (see, for example, FIG. 7). ), Measuring means (for example, standardized shoe measuring unit 112 in FIG. 5) that generates information including at least the measurement values of the plurality of items as the measurement data of the item (for example, shoe measurement data Ks1 to Ksm in FIG. 2).
Is provided.

1: shoe measuring device 2: tester foot measuring device 3: learning device 4: user foot measuring device 5: matching device 6: user terminal 11: shoe measuring DB
12: Foot measurement DB
13: Evaluation DB
21: CPU
22: ROM
23: RAM
24: Bus 25: Input / output interface 26: Output unit 27: Input unit 28: Storage unit 29: Communication unit 30: Drive 31: Removable media 111: Scan unit 112: Standardized shoe measurement unit 121: Scan unit 122: Standardized foot measurement Unit 123: Evaluation acquisition unit 141: Scan unit 142: Standardized foot measurement unit 143: Evaluation acquisition unit 151: Matching method determination unit 152: User preference acquisition unit 153: Matching unit 154: Recommendation unit J: Ball joint A: Foot circumference B : Foot width C: foot length D: heel shape E: toe height F: toe shape G: distortion H: evaluation data L: landmark point U: user

Claims (3)

Based on the measurement data of the item, identify the correct pair, detect one or more landmark points, calculate a reference symbol based on a predetermined rule for eliminating the design of the item, Measuring means for calculating each measurement value of a plurality of items based on a landmark point and the reference symbol, and generating information including at least each measurement value of the plurality of items as measurement data of the item Information processing device. An information processing method executed by an information processing apparatus,
Based on the measurement data of the item, identify the correct pair, detect one or more landmark points, calculate a reference symbol based on a predetermined rule for eliminating the design of the item, A measurement step of calculating each measurement value of a plurality of items based on a landmark point and the reference symbol, and generating information including at least each measurement value of the plurality of items as measurement data of the item Information processing method. In the computer that controls the information processing device,
Based on the measurement data of the item, identify the correct pair, detect one or more landmark points, calculate a reference symbol based on a predetermined rule for eliminating the design of the item, A measurement step of calculating each measurement value of a plurality of items based on a landmark point and the reference symbol, and generating information including at least each measurement value of the plurality of items as measurement data of the item A program that executes control processing.