Printer Friendly

The quantitative model of the product style based on eye tracking technique and Fourier decomposition.

1. Introduction

With the social development, the outlook and style of products tend to be assimilation. To enhance the brand recognition among consumers, many companies improve the identification of the product. The similarity of products' outlook plays an important role in human engineering, product function and expectation on brands (Marielle and Jan, 2005). However, although the designing approach is of great importance to meet the market demand, the decision makers always do this according to his own experience (Costa et al., 2012; Hsiao and Tsai, 2005; Zhang et al., 2013). Meanwhile, the recognition difference between users and designers on products design and the shortage of quantitative analysis of designing strategy (Shang et al., 2000) results in perceptual judgments from decision makers. Then it is hard to systematically assess the recognition of the designing scheme and product style, which makes a difference in position of designing strategies.

How can designers face the uncertainties during the process of decision making? How to make an objective and reasonable analysis and assessment on products' design? According to references (Chan, 2000; Li, 2011; Mccormack et al., 2004; Stiny and Mitchell, 1978), the style can be measured by the understanding of those unique designing elements. However, the measure method only aims at certain styles, not involves the difference of designs. Accordingly, the paper proposes the quantitative model of the product style based on eye tracking technique and Fourier decomposition, utilizes quantitative method to make calculation about product style so as to provide new ways for the designing of the company's brand strategy.

2. The Quantification of Modeling Style

The quantification of product style involves the research on influential factors of product style characteristics and its parametric processing. Then the perceptual and obscure style will be transformed to the reasonable and quantitative measured parameter.

The paper, through the eye-movement experiment, obtains characteristics that influence product style, utilizes shape grammar to extract the contour of feature elements and transforms it to two-dimensional closed curve, then makes similarity analysis on different feature elements via Fourier decomposition. After an analysis of data acquired from observers' eye movement in the regions of interest (ROI), the author makes a weight analysis on the feature elements in product modeling, then carries out a significant research on each element, finally calculates the overall similarity of different products so as to provide guidance for product design and analysis. The theoretical model is depicted above (Puello, 2016).

3. The Extraction of Style Feature Elements

When it comes to the research on style features, (Xu et al., 2007) based on the analysis of it with multidimensional scaling and morphological analysis, proposed an optimized design approach for imagery modeling on the strength of Genetic Algorithm (GA). After obtaining the data about observers' recognition on different product styles via the eye-movement experiment, the author makes a further extraction of the style features that influence the product's style.

3.1. The Design of Experiment

The paper makes a quantitative analysis on experiment objects including the headstock of CRH380A, CRH380B, CRH380C, CRH380D. The eye-movement experiment aims at identifying the characteristic elements that influence product modeling style.

In the experiment, the EysSo Ec60 remote-measuring eye tracker is the apparatus, and the experiment object is CRH380A train. To ensure that experimenters would stare at these trains at a real environment and avoid other irrelevant variables, the grey pictures presenting headstocks at a 45-degree angle with the same depth of field are chosen as experiment objects, which as the table 1 shows.

In the experiment, observers are 20 first-grade postgraduate students majoring in designing, as is shown in the table 2. They are required to judge the designing style of these trains within 20 seconds via observing the characteristic elements of the modeling.

3.2. Experimental Analysis

After the experiment is finished, the line of sight and heat maps of experiment objects are recorded as figure 2.

In the figure, the heat map displays the duration of experimenters' sight on specific objects, among which, red, yellow and green respectively stands for the longest, the second longest and the relatively short duration time, and the colorless area indicates that no one stared at this place. The line graph reveals the order of experimenters' interest in specific elements under the certain requirement, namely recognize the product style. The size of the spot represents relative length of duration of sight. To some extent, the research can be made on the characteristic elements that influence the high-speed train's modeling style according to the feature of experimenters and the graph of line of sight.

According to the result, for the experimenters, their line of sight mainly focused on the headlight, coupling gear, front windows, body contour and side windows. And the order of the line of the sight is the coupling gear, body contour, headlight, front windows and side windows. The significant research on characteristic elements for the whole product style will be made later. Therefore, the elements that influence high-speed train style were selected, including such five designing elements as coupling gear, body contour, headlight, front train window and side windows. And to simplify the experimental analysis, some small areas have been combined.

4. The Line Profile of Modeling Style

4.1. Shape Grammar

In the field of making quantitative analysis on product style, (Orsborn et al., 2006) used shape grammar to quantify the difference between cars with different kinds of rules, then designed the car within the range of parameters. It provided guidance for the design of new cars. To be more precise, it limited the design within specific range of parameters as well as combined rules and scopes via crossing the boundary of different types of cars so as to create the unique, interesting and mixed style.

The shape grammar, a way to analyze and create shape, refers to a language that can creating the two-dimensional geometric figure with a syntactic rule. Its advantage is that it can use the contour profile to show the graphic feature of the research object. The shape grammar illustrates the process of designing with the application of its rules, namely the grammar. Taking the corresponding original shape in the designing rule as an example, such rule consists of the left and the right shape, if the left shape matches the original one, then the right one will replace the original one, which is shown in the figure 3.

The limitation of the shape grammar is that the change of pictures must base on the parameters, that is to say, it requires similar shapes for any forms of change including zoom-out and rotation. To deal with that problem, the paper uses Fourier decomposition to parameterize designing elements. To meet the requirements of Fourier decomposition on closed property of the curve and avoid uncontrollable variables during the process of the quantification of designing elements, the paper, according to the shape grammar, changes the three-dimensional modeling of the high-speed train into the two-dimensional one to make analysis. Meanwhile, the paper, to some extent, simplifies the analysis of the quantification of the design.

4.2. The Contour Profile of Modeling Elements

When utilizing shape grammar to measure the product style, the designer must determine the influence of each designing element on the whole modeling. However, the designers often make judgments based on the appearance of shape elements and their own experience, recognition and emotions, which are obscure and uncertain. The paper makes an analysis on the order of line of sight and interest areas of observers via the eye-movement experiment, and then makes quantitative analysis and assessment on each factor that influence the modeling. According to the feature of high-speed train modeling, the paper chooses the front view and side view of shapes that, to a large extent, represent the product modeling as target objects, and then make quantitative analysis on modeling elements of each view. The extraction of characteristic elements from each view is shown in figure 4.

The list of feature elements of the front and side view is as followed.

4.3.The Quantitative Analysis on Modeling Elements

In the field of the quantification of designing elements, (McGarva and Mullineux, 1993) made a research on the closed planar curve that expressed by the Fourier coefficient and proposed the theoretical method that using harmonic wave to represent the closed curve to make quantitative analysis on the design. The Fourier decomposition represents the repeated periodic curves. Taking CRH380A, CRH380B, CRH380C, CRH380D as research objects, the paper extracts ten designing elements from each model to make a simple closed curve that expressed by the point set. (Orsborn et al., 2006) applies the Fourier decomposition and the genetic algorithm to design the car, the Fourier coefficient that he used can be expressed by the equation (1) below.

[a.sub.m] = [N.summation over (k=0)] ([[t.sub.k+1] - [t.sub.k]/2] [z.sub.k+1]exp(-2[pi][imt.sub.k+1])+[z.sub.k] exp(-2[pi]im[t.sub.k])) (1)

In the equation, [[alpha].sub.m] is a complex number. [Z.sub.k] represents the point in the curve, its approximate value comes from N parts of the curve. [t.sub.k] refers to the distance between point [Z.sub.k] and [Z.sub.k+1]. According to (Orsborn et al., 2006), the calculation of the distance between point and point in the equation is as followed.


The two curve can be solved by (Orsborn et al., 2006)


The value of m is close to 10. Only the first ten harmonic waves are of significance, so the distance between two curves, (k, l), can be viewed as the distance weighted sum of the first ten harmonic curves. According to (Orsborn et al., 2006), a(m) is an important weight factor for the first harmonic curve instead of the last one, because the first one is easier to observe. a(m) is expressed by the form of exponent, its approximate weight is [1.08e.sup.-0.08m] (Yannou et al., 2008). Based on the equation (1) [a.sub.m]=[u.sub.m]-[iv.sub.m], the initial distance can be calculated from the difference between [g.sub.k,m] and [g.sub.l,m], [g.sub.k,m] is the harmonic number m comes from the curve k, and [g.sub.l,m] is the harmonic number m comes from the curve l. [u.sub.max] and [u.sub.min] are the maximum and minimum value of [u.sub.k], [v.sub.max] and [v.sub.min] are the maximum and minimum value of [v.sub.k]. SimInd(k,1) is an approximate value that comes from the curve k and the curve l, and it ranges from 0% to 100%. The higher the similarity value, the more similar the curve is.

The Fourier decomposition can be viewed as a designing quantitative method to calculate the similarity value of different modeling style elements. The similarity of the body contour from the side view of 4 types high-speed train of CRH380 can be seen in the figure 5. The numerical difference of each iteration of the product style can be obtained through the comparison between the similarity value of the modeling elements of different generations, which can be the guidance for the product style iteration.

5. The Weight Analysis of Feature Elements

The first experiment focuses on the influence of feature elements on modeling style. For the further quantitative analysis on its significant influence, the author, based on the part 1 and part2, chooses the front and side view that, to a large extent, represent the style of CRH380A to carry out eye-movement experiment. Before the experiment, the author delimits the corresponding are of interest (AOI) for the specific elements. The data in the AOI can reflect the significant influence of each feature element on the whole style. The set of AOI in the eye-movement experiment is for the author to make analysis on the area that related to the experiment, the data out the region would not be counted. The result is as followed.

The statistics are as followed.

According to the above three groups of data, the order of significant feature elements is the body contour, front window, side window, coupling gear, then headlight, and their weights respectively are 46.77%, 22.30%, 15.51%, 11.68%, 3.74%.

According to the above three groups of data, the order of significant feature elements is the body contour, front window, headlight, coupling gear, side window, and their weights respectively are 46.50%, 19.04%, 14.40%, 14.59%, 5.47%.

It can be observed from the experimental result that the body contour has the largest influence on the modeling style of CRH380A, which also verifies people's real-time observation and cognitive disposition will be influenced by their personal sense. Then the second largest influential factor is the front window. Therefore, when it comes to the recognition, the designer can give prominence to such distinctive feature factor as the front window so as to differentiate it from other product styles.

6. The Similarity Analysis of Product Modeling Style

The paper extracts the five feature elements of high-speed train via the eye-movement experiment, then uses shape grammar rules to make the line profile of the product modeling to get the planar elements from the front and side view, and make the quantitative analysis on 40 design elements of four kinds of high-speed train with the Fourier decomposition. After that, the author utilizes the element weight from the eye-movement to ultimately calculate the overall similarity value. The similarity between CRH380A and CRH380B is 21.62%, and that between CRH380B and CRH380C is 30.67%. Through this way, the similarity value of different product styles can be quantified to provide guidance for the recognition strategy of product designing. Meanwhile, such method can also be used in the iteration of products, making the changes on product style of different generations at a stable level so as to such avoid strategy mistake as designing the product that not accepted by the consumer. Besides, in the paper, the significant research on feature elements also provide guidance for highlighting the brand of the product in the process of designing, receiving recognition through attracting people's line of sight and ultimately making a unique brand.

7. Conclusion

The paper puts forward a theoretical model for quantitative analysis on the product modeling style due to the perceptual judgments appeared in the process of analysis and decision making. It combines the data acquired from the eye-movement experiment to make research on the level of significance in product feature elements. Meanwhile, the author uses shape grammar rules to make two-dimensional deconstruction of product modeling, utilizes the Fourier parameter to make quantitative analysis on feature elements so as to obtain the similarity value of different product modeling style. With such theoretical method, the designer can quantify the influence level of feature elements on the whole modeling and the difference of similarity between different product modeling style so as to provide guidance for the recognition of design and highlighting the brand style.

In terms of quantification of the design, the paper uses shape grammar to make the line profile for the feature elements of product modeling, and takes the Fourier decomposition as the quantification method of feature elements, making the judgments on product designing change from the qualitative one to the quantitative one, then effectively avoids the perceptual deviation in designing. In the experiment, due to the uncontrollable factors and variables in the experimental environment, such as the light and the unbalanced move of people's eyes, there may be a little deviation in the result. Accordingly, it is necessary to pay attention to the experiment settings. And the research directions in the future are as followed. Firstly, making significant research on feature elements and the whole modeling style from different latitudes. Secondly, how can the quantitative analysis provide guidance for practice in a more effective way.

Recebido/Submission: 11/04/2016

Aceitacao/Acceptance: 25/07/2016


Chan C. S. (2000). Can style be measured? Design Studies, 21(3), 277-291.

Costa C.A., Luciano M.A., Lima C.P., Young R.I.M. (2012). Assessment of a product range model concept to support design reuse using rule based systems and case based reasoning. Advanced Engineering Informatics, 26(2), 292-305

Hsiao S.W., Tsai H.C. (2005). Applying a hybrid approach based on fuzzy neural network and genetic algorithm to product form design. International Journal of Industrial Ergonomics, 35(5), 411-428.

Li I.K. (2011). Computing style. Nexus Network Journal, 13(1), 183-193. DOI: 10.1007/ s00004-011-0053-9

Marielle E.H. Creusen, Jan P.L. Schoormans. (2005). The Different Roles of Product Appearance in Consumer Choice. Journal of Product Innovation Management, 22(1), 63-81. DOI: 10.1111/j.0737-6782.2005.00103.x

Mccormack J. P., Cagan J., Vogel C. M. (2004). Speaking the buick language: capturing, understanding, and exploring brand identity with shape grammars. Design Studies, 25(1), 1-29.

Mcgarva J., Mullineux G. (1993). Harmonic representation of closed curves. Applied Mathematical Modelling, 17(4), 213-218. DOI: 10.1016/0307-904X (93) 90109-T

Puello, Jose de J. Jimenez; Feliu, Tomas San; Calvo-Manzano, Jose A. (2016). Una aproximacion basada en metamodelado del area de proceso de Validacion del CMMI: Un caso de estudio. RISTI - Revista Iberica de Sistemas e Tecnologias de Informacao, 17, 26-40.

S O., J C., R P., Rc S. (2006). Creating cross-over vehicles: defining and combining vehicle classes, using shape grammars. Artificial Intelligence for Engineering Design Analysis & Manufacturing, 20(3), 217-246, DOI: S0890060406060185

Shang H.H., Ming C.C., Chang C.C. (2000). A semantic differential study of designers' and users' product form perception. International Journal of Industrial Ergonomics, 25(4), 375-391.

Stiny G., Mitchell W. J. (1978). The palladian grammar. Environment & Planning B Planning & Design, 5(1), 5-18, DOI: 10.1068/b050005

Xu J., Song S.Q., Zhang K.J. (2007). Product image form optimization design based on genetical gorithm. Journal of Mechanical Engineering, 43(4), 53-56.

Yannou B., Dihlmann M., Cluzel F. (2008). Indirect encoding of the genes of a closed curve for interactively create innovative car silhouettes. In: Proceedings of theioth International Design Conference-DESIGN2008, 1243-1254.

Zhang Y., Luo X., Li J., Buis J.J. (2013). A semantic representation model for design rationale of products. Advanced Engineering Informatics, 27(1), 13-26.

Guoshu Yuan, Wenzhe Cun, Qingsheng Xie, Jian Lyu, Weijie Pan,,,,

Key Laboratory of Advanced Manufacturing Technology, Ministry of Education, Guizhou University, Guiyang, 550003, China

Table 2--Experimental population setting

serial     experimental            profession           number
number      population                                 of people

A        graduate student         design major             5
B        graduate student   mechanical related major       5
C        College Student           unlimited               5
D           unlimited              unlimited               5

serial   average        gender           corrected     blind
number     age                         visual acuity   seruo

A          24      3 men and 2 women       good         no
B          23      3 men and 2 women       good         no
C          22      4 men and 1 women       good         no
D          32      2 men and 3 women       good         no

Table 3--List of feature elements of CRH380A type high speed train

view           elements that affect the style

front   Front     front       front    coupler   side
view    profile   headlight   window             window

side    Side      side        front    coupler   side
view    profile   headlight   window             window

Table 4--The AOI data analysis of Side profile (unit: ms)

                            AOI1     AOI2    AOI3    AOI4    AOI5

fixation time               149347   71227   49572   37322   11955
number of fixation points   455      213     174     43      44
average fixation duration   336      334     285     868     273

AOI1, Side profile; AOI2, front window; AOI3, side window; AOI4,
coupler; AOI5, side headlight;

Table 5--The AOI data analysis of front profile (unit: ms)

                            AOI1     AOI2    AOI3    AOI4    AOI5

fixation time               166459   68161   51565   52230   19577
number of fixation points   510      199     139     111     49
average fixation duration   326      343     371     471     400

AOI1, front profile; AOI2, front window; AOI3, coupler; AOI4, front
headlight; AOI5, side window;

Agricultural University of Hebei, Baoding, Hebei 071001, China
COPYRIGHT 2016 AISTI (Iberian Association for Information Systems and Technologies)
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2016 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Yuan, Guoshu; Cun, Wenzhe; Xie, Qingsheng; Lyu, Jian; Pan, Weijie
Publication:RISTI (Revista Iberica de Sistemas e Tecnologias de Informacao)
Date:Oct 15, 2016
Previous Article:Study on "people-oriented" college students management.
Next Article:On teaching reform of college English based on mobile learning.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |