Professional Documents
Culture Documents
A Systematic Approach in Shoe Last Design For Human Feet: C. S. Wang, T. R. Chang, M. C. Lin
A Systematic Approach in Shoe Last Design For Human Feet: C. S. Wang, T. R. Chang, M. C. Lin
II. METHODOLOGY
Abstract - The main purpose of this research is to build a
process to find the most suitable shoe last for making shoes for The purpose of this research is to obtain the point
human feet. A fitness function is defined to determine the clouds data of human feet and shoe lasts in the reverse
optimum shoe last. Based on reverse engineering (RE) engineering method by CMM and to search the coordinate
technology, this research scanned the surface of human foot and data of the features of girths on the surface. Subsequently,
shoe last in STL (Stereo Lithography) format. We proposed the fuzzy theory and AHP will be applied to find the optimal
STL feature-based slicing algorithm to automatically construct
the three most important girth characteristics for human feet and
shoe last by analyzing the data to describe the
shoe last. Fuzzy theory was used to analyze and build the characteristics of human feet and shoe lasts. The primary
membership functions of these three important characteristics methods and the approach are described below (Fig. 1):
between the shoe last and human feet. The analytical hierarchy 1) Scan human feet and shoe lasts to obtain surface
process (AHP) was applied to compare the important index and and point cloud data [4, 5, 6].
decide the weighting functions for each girth characteristic to 2) Compute and figure out the three important girths
determine the fitness function in all shoe last databases for the using a triangular, STL-slicing algorithm.
feet. 3) Determine the membership functions in shoe last
girth features by fuzzy theory.
Keywords - Reverse Engineering; Anthropometry; Shoe
4) Find the weights for each girth by AHP analyzing
Last Design; STL Feature-based; Slicing Process; Fuzzy
Theory; Analytical Hierarchy Process
process.
5) Case study and rank to find the most suitable
I. INTRODUCTION shoe last for the human feet.
Last and Human Feet
Pairwise Comparison
Samples
The shoe last is designed according to the features Matrix of the Girth by
Questionnaires
and characteristics of human feet and is concerned with
Getting 3D
it’s the shape and the functions of feet. The most appearance data
Outline Data
by CMM
important features of feet that the shoe last must take into Find the Eigen
account are length, width, ball girth, instep girth, and Characteristics by
Vector for the Matrix
waist girth. Length and width are related to the size of the Slicing Process
shoe last, and the three girths are related to comfort and The Weights of
the Girth
fitness when people are wearing the shoes. The main Constructing
Characteristics Data Characteristics
purpose of this research is to design a process to find the
most suitable shoe last in the database of shoe lasts for the
human feet based on the length, width, and three main Fuzzy Analyzing and Ranking through
girths of shoe last. An STL (Stereo Lithography), feature- the Girth Characteristics and the
Weights
based, slicing algorithm was used to automatically
construct the three most important feature-girth The Optimal Shoe Last for the
characteristics of human feet and shoe last. A fitness Human Feet
⎪( LR − x ) /( LR − LG ), LG ≤ x ≤ LR .......μB
Fig. 6. Intersection point ⎪⎩0, x > LR
III. RESULTS
4 (ab8562)
The user interface that implements the method
mentioned above and obtains the measurements of the
characteristic features is programmed by MATLAB, as
shown in Fig. 8. The girths can be sent to Geomagic 5 (ab9647)
Studio or AutoCAD formats for further analysis and
visualization, as shown in Fig. 9 and Fig. 10.
6 (tw698)
7 (tw736)
8 (tw9615)
9 (tw1127)
10 (tw1247)
Fig. 9. Output the girth to Geomagic for further analysis
Although the ranking of the optimal shoe last can be The concrete efforts in this paper that have been
obtained by the estimation obtained from the proposed completed are listed below:
method using fuzzy theory and AHP, the shoe last that fits 1. A programmed graphic user interface, which
each examinee the best can be determined according by automatically estimates and obtains the data
the ranking. Once the database that has considerable data characteristics of the shoe lasts and human feet, has
related to shoe last design is set up with the increase of been created, allowing human errors to be decreased.
the lasts, there will be a need for faster ways to obtain a 2. A concept of the sense threshold applying to shoe last
more accurate ranking among all shoe lasts and to and feet from fuzzy theory and the weights of girths
determine the optimal shoe last for each person. by AHP is proposed. A fitness function in shoe last
and feet is proposed, and the optimal shoe last for
1) Case 1: First Examinee individual feet can be found. All the shoe lasts for
There are no shoe lasts for which all the membership one examinee can be ranked by using the fitness
grades of these three girths are not 0 among these 10 lasts, function.
so there was no optimal shoe last for first examinee
among these 10 lasts. However, for Nos. ab9998, ab8652, 3. All the characteristics and the graphic data of the
and tw9615, these three shoe lasts all have only two girth shoe lasts and human feet estimated within this
items for which the membership grades are not 0. research can be stored to set up a database that can be
Therefore, the optimal shoe last among these three lasts used in subsequent work as a reference for the
can be estimated through the accurate ranking of the modification or design of shoe lasts.
formula, R = wB M B + wI MI + wW MW , and the result is
shown below:
Rab9998 = 0.9772 × 0.671 + 0.0816 × 0.885 + 0.1848 × 0 = 0.7279 REFERENCES
Rab8652 = 0.9772 × 0.648 + 0.0816 × 0 + 0.1848 × 0.315 = 0.6914 [1] Choi, S.H., and Kwok, K.T., 2002, “Hierarchical slice
Rtw9615 = 0.9772 × 0 + 0.0816 × 0.523 + 0.1848 × 0.307 = 0.0994 contours for layered-manufacturing”, Computers in
Industry, Vol. 48, pp. 219-239.
As a result, Rab9998 > Rab8652 > Rtw9615 , which means [2] Douligeris, C., and Pereira, I., 1992, “An analytical
hierarchy process approach to the analysis of quality in
the No. ab9998 shoe last would be the optimal last for the telecommunication systems”, IEEE Proceedings on SMC,
first examinee among these 10 lasts, and it is shown in pp. 1684-1688.
Fig. 11(a) for STL and in Fig. 11(b) for the sample. [3] Gibson, I., 2002, Software solutions for rapid prototyping,
2) Case 2 - Second Examinee Ch 5, John Wiley.
The No. ab9998 last has all the membership grades that [4] Kouchi, M., 1995, “Analysis of foot shape variation based
are not 0 for these three girth items. Therefore, it would on the medial axis of foot outline”, Ergonomics, Vol.38,
No.9, pp.1911-1920.
be the optimal last for the second examinee among these
[5] Kouchi, M., and Tsutsumi, E., 1996, “Relation between
10 lasts, and it is shown in Fig. 11(a) for STL and in Fig. the medial axis of the foot outline and 3-D foot shape”,
11(b) for the sample. Ergonomics, Vol. 39, No. 6, pp.853-861.
3) Case 3 - Third Examinee [6] Kouchi, M., and Mochimaru, M., 2001, “Development of
Last No. tw736 is the only one for which all the low cost foot-scanner for a custom shoe making system”,
membership grades of these three girth items are not 0. Proceeding of 5th Symp. of Footwear Biomechanics, pp.
Therefore, it would be the optimal last for the third 58-59.
examinee among these 10 lasts, and it is shown in Fig. [7] Saaty, T.L., 1980, “The Analytic Hierarchy Process”,
12(a) and Fig. 12(b). McGraw-Hill, New York.
[8] Zadeh, L.A., 1965, “Fuzzy sets”, Information and
Control, Vol. 8, pp. 338-353.