CN104921410B - A kind of shoe tree parameter automatic prediction method and prediction meanss based on dual model - Google Patents
A kind of shoe tree parameter automatic prediction method and prediction meanss based on dual model Download PDFInfo
- Publication number
- CN104921410B CN104921410B CN201510405858.0A CN201510405858A CN104921410B CN 104921410 B CN104921410 B CN 104921410B CN 201510405858 A CN201510405858 A CN 201510405858A CN 104921410 B CN104921410 B CN 104921410B
- Authority
- CN
- China
- Prior art keywords
- foot
- shoe tree
- model
- shoe
- parameters
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000009977 dual effect Effects 0.000 title claims abstract description 18
- 238000000034 method Methods 0.000 title claims abstract description 17
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000013145 classification model Methods 0.000 claims description 26
- 238000012706 support-vector machine Methods 0.000 claims description 19
- 210000002683 foot Anatomy 0.000 description 79
- 238000013461 design Methods 0.000 description 7
- 210000003813 thumb Anatomy 0.000 description 6
- 210000000454 fifth toe Anatomy 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 210000003371 toe Anatomy 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Landscapes
- Footwear And Its Accessory, Manufacturing Method And Apparatuses (AREA)
Abstract
The present invention relates to a kind of shoe tree parameter automatic prediction method based on dual model, it is characterised in that comprise the following steps:(S1) training of foot and last carving disaggregated model, foot and last carving difference model;(S2) dual model automatic Prediction shoe tree parameter is utilized.In addition, the invention further relates to a kind of shoe tree parameter automatic Prediction device based on dual model, including data acquisition module, for gathering foot parameter;Dual model training module, for training foot with shoe tree disaggregated model and foot and shoe tree difference model;Memory module, for preserving foot data, basic shoe tree parameter, foot and shoe tree disaggregated model, foot and shoe tree difference model;Automatic Prediction module, for automatic Prediction shoe tree parameter.For the strong problem of current shoe last designing inefficiency, subjectivity, the general character of client's wear shoes is excavated, the shoe tree parameter automatic prediction method and device based on dual model is designed, according to the foot parameter automatic Prediction of client and suitable shoe tree parameter can be designed.
Description
Technical Field
The invention relates to a parameter automatic prediction method and a prediction device, in particular to a shoe tree parameter automatic prediction method and a prediction device based on double models.
Background
With the progress of society, the development of economy, the improvement of people's living standard, people are also improving gradually to the style of shoes, comfort level and the requirement of shoemaking technology, and common shoes are the code sharing mostly in the market, and according to empirical design, to most consumers, have the risk of not fitting the foot, and the customization pair of shoes that fit the foot, not only time-consuming, the cost is too expensive, and ordinary consumer often can't bear.
The shoe tree is a reference for manufacturing the shoes, has direct relation with comfort level, style and the like of the shoes, and is a key node for restricting consumers from customizing proper shoes. At present, shoe tree parameters are generally obtained by correcting basic foot shape parameters through an experienced technician, foot shape data are formed by some basic characteristic parameters obtained manually, correction and transition of the shoe tree at the characteristic points are mostly determined by experience of a designer, and the shoe tree model design method has strong subjectivity, low efficiency and high price, so that the requirement of quickly designing a high-quality shoe tree is difficult to meet through the traditional shoe tree model design method, and design and manufacture of comfortable, high-quality and low-price shoes are restricted.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention aims to provide a method and a device for automatically predicting shoe tree parameters, which can be used for rapidly and accurately predicting the shoe tree parameters.
In order to achieve the aim, the invention provides a shoe tree parameter automatic prediction method based on double models, which comprises the following steps:
step one, training a foot-last classification model and a foot-last difference value model;
and step two, automatically predicting the parameters of the shoe tree by using the double models.
Preferably, the step one further comprises the steps of:
step S101, measuring foot parameters of a tester;
s102, enabling a tester to try on the shoes made of the basic shoe trees and matching the most suitable shoe size;
step S103, taking different shoe sizes as different categories and foot parameters as characteristics, and training a classifier by using an SVM (support vector machine) to obtain a foot and last classification model;
step S104, reading in basic shoe tree parameters, corresponding to the foot parameters one by one, and making difference values;
and S105, counting the expected difference value of the same shoe size to obtain a difference value model of the foot and the last.
Preferably, the second step further comprises the steps of:
step S201, measuring foot parameters of a consumer;
step S202, predicting the number of shoe tree codes by utilizing a foot and shoe tree classification model;
step S203, obtaining the difference value of the foot and the last corresponding to the number of the last codes in the difference value model of the foot and the last;
and step S204, adding the corresponding difference value to the foot parameters to obtain the parameters of the final last.
The invention also provides a shoe tree parameter automatic prediction device based on the dual model, which comprises a data acquisition module, a parameter prediction module and a parameter prediction module, wherein the data acquisition module is used for acquiring foot parameters; the double-model training module is used for training a foot-to-shoe tree classification model and a foot-to-shoe tree difference value model; the storage module is used for storing foot data, basic shoe tree parameters, a foot-shoe tree classification model and a foot-shoe tree difference value model; and the automatic prediction module is used for automatically predicting the parameters of the shoe tree.
Aiming at the problems of low design efficiency and strong subjectivity of the current shoe tree, the method and the device for automatically predicting the shoe tree parameters based on the dual models are designed by mining the commonality of shoes worn by customers, and the proper shoe tree parameters can be automatically predicted and designed according to the foot parameters of the customers.
Drawings
FIG. 1 is a flow chart of a first preferred embodiment of the present invention;
FIG. 2 is a flow chart of a second preferred embodiment of the present invention;
fig. 3 is a schematic structural diagram of a prediction apparatus provided in the present invention.
Detailed Description
As shown in FIG. 1, the invention provides a shoe tree parameter automatic prediction method based on dual models, which comprises the following steps:
step one, training a foot-last classification model and a foot-last difference value model;
and step two, automatically predicting the parameters of the shoe tree by using the double models.
Advantageously, said step one further comprises the steps of:
s101, measuring foot parameters of a tester;
s102, enabling a tester to try on the shoes made of the basic shoe trees and match the most appropriate shoe size;
s103, taking different shoe sizes as different categories and foot parameters as characteristics, and training a classifier by using an SVM (support vector machine) to obtain a foot and last classification model;
s104, reading in basic shoe tree parameters, corresponding to the foot parameters one by one, and making difference values;
and S105, counting the expected difference value of the same shoe size to obtain a difference value model of the foot and the last.
As shown in fig. 2, the first step may further include the following steps:
s111, measuring foot parameters of the testee;
s112, enabling a tester to try on the shoes made of the basic shoe trees and match the most suitable shoe size;
s113, reading in basic shoe tree parameters, corresponding to the foot parameters one by one, and making difference values
S114, counting the expected difference value of the same shoe size to obtain a difference value model of the foot and the last;
and S115, taking different shoe sizes as different categories and foot parameters as characteristics, and training a classifier by using an SVM (support vector machine) to obtain a foot and last classification model.
In the step S101, when measuring the foot parameters of the tester, the parameters specifically include the foot length, the back length, the heel margin, the plantar circumference G2, the plantar circumference G3, the heel circumference G4, the foot width, the palm width, the inner width of the outline of the thumb protrusion point 5, the outer width of the outline of the little toe protrusion point, the inner width of the outline of the first metatarsophalangeal toe, the outer width of the outline of the fifth metatarsophalangeal toe, the outer width of the lumbar fossa outline, the full width of the heel center wheel outline, the height of the thumb, the height of the foot shape G2, the height of the foot shape G3, the height of the foot shape G4, the projected foot bottom area, the area of G2, the area of G3, the area of G4, the front middle section area of G4, the volume of the front of.
In the step S102, the tester tries to wear the shoes made by the basic shoe last to match the most suitable shoe size 0, and here, a plurality of testers need to try to wear the shoes made by the basic shoe last, but when the shoes are matched to the most suitable shoe size, the shoe size is recorded and is associated with the foot parameters of the testers to obtain the classification model.
In the above steps S103 and S115, when the classification model of the foot and the shoe tree is obtained, an SVM is required to train the classifier. The SVM is called support vector machine in English, Chinese is called a support vector machine, and the SVM is a two-class classification model in popular terms, a basic model of the SVM is defined as a linear classifier with the maximum interval of 5 on a feature space, a learning strategy of the SVM is interval maximization, and the SVM can be finally converted into the solution of a convex quadratic programming problem.
In the above steps S105 and S114, when obtaining the difference value model between the foot and the last, the formula to be used is as follows:
wherein i represents different shoe sizes, Ni is the number of learning samples corresponding to the shoe size i, and C is the difference value between the foot parameter and the shoe last parameter.
Two models, namely a classification model and a difference model, can be obtained through the first step.
The second step may further comprise the steps of:
s201, measuring foot parameters of a consumer;
s202, predicting the number of shoe tree codes by utilizing a foot and shoe tree classification model;
s203, obtaining the difference value of the foot and the last corresponding to the number of the last in the difference value model of the foot and the last;
and S204, adding the corresponding difference value to the foot parameters to obtain the parameters of the final last.
In step S201, foot parameters of a specific customer who wants to customize a shoe are measured, specifically, the parameters include foot length, back length, heel margin, plantar circumference G2, plantar circumference G3, pocket heel circumference G4, foot width, foot sole width, thumb protrusion outline inner width, little toe protrusion outline outer width, first metatarsophalangeal outline inner width, fifth metatarsophalangeal outline outer width, fossa outline outer width, heel contour full width, thumb height, foot shape G2 height, foot shape G3 height, foot shape G4 height, foot projection base area, G2 area, G3 area, G4 area, G4 front middle cross-sectional area, foot front volume, and foot volume.
The foot parameters of the consumer are applied to the foot and shoe tree classification model to predict the number of the shoe tree codes, and if the difference between the predicted number of the shoe tree codes and the ordinary number of the shoe tree codes is too large, the number of the shoe tree codes is artificially corrected.
And obtaining the difference value of the foot and the shoe tree corresponding to the number of the shoe tree of the consumer by using the foot and shoe difference value model.
And adding the corresponding difference value to the foot parameters of the consumer to obtain the parameters of the final shoe tree. For example, the length of the last can be obtained by adding the value of the length of the consumer's foot to the difference corresponding to the length of the foot obtained from the difference model.
Advantageously, in step two, the obtained parameters of the final last are written into a database for incremental learning of the model.
As shown in fig. 3, the present invention further provides a dual model-based shoe tree parameter automatic prediction apparatus, which comprises a data acquisition module for acquiring foot parameters; the double-model training module is used for training a foot-to-shoe tree classification model and a foot-to-shoe tree difference value model; the storage module is used for storing foot data, basic shoe tree parameters, a foot-shoe tree classification model and a foot-shoe tree difference value model; and the automatic prediction module is used for automatically predicting the parameters of the shoe tree.
The data acquisition module is used for acquiring foot parameters, and specifically acquiring parameter information including foot length, back length, heel margin, foot plantar circumference G2, foot plantar circumference G3, heel circumference G4, foot width, palm width, inner width of a thumb outer protruding point contour, outer width of a little toe outer protruding point contour, inner width of a first metatarsophalangeal contour, outer width of a fifth metatarsophalangeal contour, outer width of a lumbar fossa contour, full width of a heel center contour, thumb height, foot type G2 height, foot type G3 height, foot type G4 height, foot projection base area, G2 area, G3 area, G4 area, G4 front middle cross-sectional area, foot front volume and whole foot volume.
Preferably, the dual model training module includes a foot-to-last classification model training unit and a foot-to-last difference model training unit. Wherein, the foot and shoe tree classification model training unit uses SVM to train the classifier. The shoe tree difference model training unit uses the following formula during working:
wherein i represents different shoe sizes, Ni is the number of learning samples corresponding to the shoe size i, and C is the difference value between the foot parameter and the shoe tree parameter.
Advantageously, the automatic prediction module comprises a last code number prediction unit for applying the foot parameters of the consumer to the foot and last classification model to predict the last code number, and if the difference between the predicted last code number and the usual code number is too large, the predicted last code number is artificially corrected.
Preferably, the automatic prediction module includes a foot-last difference obtaining unit, configured to obtain a difference between a foot and a last corresponding to the number of the customer's last codes.
Aiming at the problems of low design efficiency and strong subjectivity of the current shoe tree, the method and the device for automatically predicting the shoe tree parameters based on the dual models are designed by mining the commonality of shoes worn by customers, and the proper shoe tree parameters can be automatically predicted and designed according to the foot parameters of the customers.
The foregoing has described in detail preferred embodiments of the present invention. It should be understood that numerous modifications and variations can be devised by those skilled in the art in light of the above teachings. Therefore, the technical solutions available to those skilled in the art through logic analysis, reasoning and limited experiments based on the design concept of the present invention should be within the scope of the present invention and/or the protection scope defined by the claims.
Claims (8)
1. A shoe tree parameter automatic prediction method based on double models is characterized by comprising the following steps:
step one, training a foot-last classification model and a foot-last difference value model;
step two, automatically predicting shoe tree parameters by using the double models;
wherein,
the first step further comprises the step of,
(S101) measuring a subject' S foot parameters;
(S102) enabling a tester to try on the shoes made of the basic shoe trees and matching the most suitable shoe size;
(S103) taking different shoe sizes as different categories and foot parameters as characteristics, and training a classifier by using an SVM (support vector machine) to obtain a foot and last classification model;
(S104) reading in basic shoe tree parameters, corresponding to the foot parameters one by one, and making difference values;
(S105) counting the expected difference value of the same shoe size to obtain a difference value model of the foot and the last;
or,
the first step further comprises the following steps:
(S111) measuring a subject' S foot parameters;
(S112) fitting the shoe made by the basic shoe tree with the most suitable shoe size by the tester;
(S113) reading in basic shoe tree parameters, corresponding to the foot parameters one by one, and making difference values;
(S114) counting the expected difference value of the same shoe size to obtain a difference value model of the foot and the last;
(S115) taking different shoe sizes as different categories and taking foot parameters as characteristics, and training a classifier by using an SVM (support vector machine) to obtain a foot and last classification model.
2. The dual model based automatic shoe tree parameter prediction method according to claim 1, wherein said second step further comprises the steps of:
(S201) measuring a foot parameter of the consumer;
(S202) predicting the number of the shoe tree codes by using the foot and shoe tree classification model;
(S203) obtaining the difference value of the foot and the last corresponding to the number of the last codes in the foot and last difference value model;
and (S204) adding the corresponding difference value to the foot parameters to obtain the parameters of the final last.
3. The dual model-based automatic shoe last parameter prediction method according to claim 2, wherein said second step further comprises the steps of:
(S205) if the difference between the predicted number of the last codes and the ordinary number of the last codes is too large, the number of the last codes is corrected manually.
4. The dual model-based automatic shoe last parameter prediction method according to claim 3, wherein said second step further comprises the steps of:
(S206) writing the obtained parameters of the final shoe tree into a database for incremental learning of the model.
5. The utility model provides a shoe tree parameter automatic prediction device based on bimodulus, its characterized in that: the device comprises a data acquisition module, a data processing module and a data processing module, wherein the data acquisition module is used for acquiring foot parameters; the double-model training module is used for training a foot-to-shoe tree classification model and a foot-to-shoe tree difference value model; the storage module is used for storing foot data, basic shoe tree parameters, a foot-shoe tree classification model and a foot-shoe tree difference value model; and the automatic prediction module is used for automatically predicting the parameters of the shoe tree.
6. The dual model-based automatic shoe tree parameter prediction device of claim 5, wherein: the double-model training module comprises a foot and shoe tree classification model training unit and a foot and shoe tree difference value model training unit.
7. The dual model-based automatic shoe tree parameter prediction device of claim 6, wherein: the automatic prediction module comprises a shoe tree code number prediction unit.
8. The dual model-based automatic shoe tree parameter prediction device of claim 7, wherein: the automatic prediction module comprises a foot and shoe tree difference value acquisition unit.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510405858.0A CN104921410B (en) | 2015-07-10 | 2015-07-10 | A kind of shoe tree parameter automatic prediction method and prediction meanss based on dual model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510405858.0A CN104921410B (en) | 2015-07-10 | 2015-07-10 | A kind of shoe tree parameter automatic prediction method and prediction meanss based on dual model |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104921410A CN104921410A (en) | 2015-09-23 |
CN104921410B true CN104921410B (en) | 2017-03-15 |
Family
ID=54109115
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510405858.0A Active CN104921410B (en) | 2015-07-10 | 2015-07-10 | A kind of shoe tree parameter automatic prediction method and prediction meanss based on dual model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104921410B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106937774B (en) * | 2017-03-30 | 2022-05-20 | 李文谦 | Shoe tree size prediction method and device based on deep learning |
CN108960262B (en) * | 2017-05-19 | 2024-02-09 | 意礴数字科技(深圳)有限公司 | Method, device and system for predicting shoe codes and computer readable storage medium |
CN107259711A (en) * | 2017-07-25 | 2017-10-20 | 余友和 | A kind of shoes match somebody with somebody code system |
CN107609933B (en) * | 2017-08-24 | 2020-06-23 | 深圳市云智数据服务有限公司 | Footage customization method and system based on store transaction records |
CN108765059B (en) * | 2018-04-28 | 2021-09-28 | 北京知足科技有限公司 | Shoe pushing method and device |
CN109086294A (en) * | 2018-06-13 | 2018-12-25 | 东莞时谛智能科技有限公司 | A kind of shoe tree database sharing and management method and system |
CN109064260B (en) * | 2018-07-11 | 2022-03-01 | 北京知足科技有限公司 | Shoe type data acquisition method and device |
CN109003167B (en) * | 2018-08-01 | 2020-12-29 | 深圳市云智数据服务有限公司 | Data processing method and device for shoe and boot customization |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4817222A (en) * | 1987-10-15 | 1989-04-04 | Shafir Aaron | Method and apparatus for making shoe lasts and/or shoe components |
CN101088728A (en) * | 2007-07-19 | 2007-12-19 | 清华大学 | Digitalized sneakers tree customizing production apparatus and process |
CN101658347B (en) * | 2009-09-24 | 2011-11-30 | 浙江大学 | Method for obtaining dynamic shape of foot model |
CN102214250B (en) * | 2010-04-12 | 2013-06-05 | 温州大学 | Method for generating statistics deformation model from shoe tree sample set |
CN102783767B (en) * | 2011-05-18 | 2014-12-17 | 香港纺织及成衣研发中心有限公司 | System for designing customized shoe tree and evaluating comfort thereof |
CN204838260U (en) * | 2015-07-10 | 2015-12-09 | 李文谦 | Automatic prediction unit of shoe tree parameter based on bimodulus type |
-
2015
- 2015-07-10 CN CN201510405858.0A patent/CN104921410B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN104921410A (en) | 2015-09-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104921410B (en) | A kind of shoe tree parameter automatic prediction method and prediction meanss based on dual model | |
EP3649883B1 (en) | A method for manufacturing an outsole and a method for generating a visual ousole pattern | |
US7409256B2 (en) | Footwear measurement and footwear manufacture systems and methods | |
US20180300791A1 (en) | Method and system for recommending fitting footwear | |
CN103646408A (en) | Image based human foot key part data extraction and measurement method | |
CN107928006A (en) | The design method and relevant apparatus of a kind of shoe last model | |
CN114041659B (en) | Shoe and boot customization method based on fitting standardization | |
CN113679147A (en) | Method for customizing shoe | |
JP2023052233A (en) | System and method for controlling shoe part production machine | |
CN106937774B (en) | Shoe tree size prediction method and device based on deep learning | |
CN204838260U (en) | Automatic prediction unit of shoe tree parameter based on bimodulus type | |
CN108596704A (en) | Large-scale personalized foot last matching method and system for customization | |
CN107609933A (en) | A kind of foot-measuring method for customizing based on shops's conclusion of the business record, system | |
CN109003167B (en) | Data processing method and device for shoe and boot customization | |
CN115081042A (en) | Rendering method, device and equipment based on sole characteristic data and storage medium | |
CN111354463B (en) | Human health measurement method, device, computer equipment and storage medium | |
CN106971424A (en) | A kind of shoe tree shoe last variation based on mesh deformation technique | |
CN103955715B (en) | Automatic person identification method based on barefoot or stockinged footprint image | |
CN109948243A (en) | A kind of method and system using foot's three-dimensional modeling data storehouse data customization insole | |
CN102063533A (en) | Method for designing shoe tree according to pelma image analysis data | |
CN104939437A (en) | Three-dimensional shoe tree model structure designing method based on different foot types | |
CN113554489A (en) | High-fit shoe customization process | |
CN115067627B (en) | Automatic customized shoe tree generation system and method, equipment, storage medium and cloud platform | |
CN102008152A (en) | Digital standard sample last | |
KR101223238B1 (en) | Method of manufacturing an insole |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |