The influence of mascot's modeling characteristics on emotional image

. We intends to investigate the impact of the characteristics of mascots modelling on the emotional imagery based on the theory of perceptual engineering. Specifically, we used the method of semantic difference image cognitive analysis on representative samples by virtue of SPSS software, which completes principal component analysis and regression analysis, typical vocabulary screening, as well as extraction of projects and category of the project modelling using morphological analysis in combination with the theory of science I analysis the correlation coefficient. Eventually, we will translate user's perceptual demand description into rational modelling characteristics, so as to acquire a better understanding of the correlation between mascots modelling characteristic and emotional image, all of which provide certain reference basis for designers.


Introduction
Considered as an important bearing way of design in modern society, mascots have been paid more and more attention by people. Excellent mascot design is recorded in people's mind for a long time with its distinctive shape and vivid image, and it has become the brand and corporate image symbol to be remembered. Study of the mascot in the design market are stays at the level of the design on the surface, with the number of mascots set to feature emotional study remaining relatively small. Moreover, the current research is targeted to theoretical narration with strong subjectivity, and some researchers design experiments to explore the emotions expressed by mascots, but their analysis stays at the level of qualitative analysis, lacking in quantitative analysis shown by experimental data and emotional tendency, the inner link of which deserves further exploration. Based on Kansei engineering, this paper constructs a correlation model between perceptual images and modeling elements to provide theoretical support for mascot modeling design, thus effectively guiding the innovative design of mascots, and thereby win better brand communication effect.
This paper takes the mascot design as the research object and quantifies the type I theory as the principle to study the correlation between the perceptual image of the mascot and its modeling elements, so as to endow the product with a special style and form, which eventually satisfy the psychological needs of contemporary consumers.

Kansei engineering
Perceptual (Kansei) originated from Japanese the word " カ ン セ ィ " transliteration, it refers to the psychological feelings of consumers and product image cognition. Combined with the engineering technology, perceptual technology is a kind of people-oriented engineering technology, it adopt the method of engineering human feelings and quantify people's emotional information so as to find out the relationship between the perceptual and the engineering quantity, and it translate the perceptual information into the design of the specific form, in order to develop and design products subject in line with the consumer demand.
In daily life, when people observe objective things around them, most of them will have certain psychological activities, and the feelings brought by such activities directly affect the satisfaction of people's inner needs when they understand things. The modeling features of the mascot design are the carrier of the audience's emotional cognition of the mascot. Through the analysis, observation, understanding of the design needs and the application of this in the designing, the designer creates the products that meet people's needs and expectations, and gives people a sense of beauty. Therefore, the basic appeal of design is emotion, only the psychological needs of consumers and use of the feeling of factors combined in the design of each step and link, can designers really effectively solve the design problem. It is the research paradigm of Kansei engineering that all kinds of emotions and feelings brought by products through semantics can be expressed. Mascot categories are numerous and complex, usually with animals or plants as the form of divergence. In order to get a better quantitative analysis, we choose an animal prototype for research. As a common mascot prototype, panda is unique and symbolic enough to convey brand awareness. After many investigations, the mascot with panda as the prototype is selected for emotional research.
A total of 46 images were collected by designing websites, periodicals, magazines, etc, and 21 mascot samples were selected after excluding samples that are too homogeneous or fuzzy. Under the condition of studying the modeling features, the interference factors such as color and volume are excluded, the image is processed in grayscale, and the brightness is adjusted to better maintain the visual consistency of the samples, and the general shape is classified and discharged, so as to facilitate the later modeling extraction [1], as shown in figure 2.

construct perceptual semantic space
Through brainstorming and network, magazine and dictionaries, dictionaries and so on, we widely collected 40 adjectives for the study, and we kept 10 groups of them, after preliminary screening. Using the Likert scale method, we write the words of opposite meaning on the two ends of the scale, which shows 7 grades of intensity for evaluation, and the middle value is set up as 0, the left and right extreme values are respectively 3 and -3. At the same time, samples numbers will be labeled in front of the table [2] . Designers engaged in design-related industries were invited to evaluate and score the samples. The subjects expressed their feelings in a quantitative form on the attitude scale of this particular descriptive word, as shown in Table 1. The mean value of perceptual images was obtained by statistical score calculation, as shown in Table. 2.  factors [1] . The positive factor load represents the positive evaluation, while the negative factor load represents the negative evaluation. After the factor analysis, the cumulative contribution value of the three factors is obtained. After the rotation matrix, the first type of factors are simple, round and advanced. The second kind of factor has the individuality, the creativity; Traditional and affinity, as shown in figure 4. Finally, the three factors advanced, creative and traditional are extracted for the later correlation analysis, and the cumulative contribution value of the three factors is obtained after factor analysis [3] . In order to make the perceptual image more representative, clustering analysis was carried out on the eight adjectives and combined it with data from factor calculation, as shown in figure 5. The distance of each image context from its group of center is calculated by coordinates of each image, and the shortest one among them is the representative of the group of image semantics, lastly, the next pairs of image semantics used in the next stage of experiment namely advanced, individuality, traditional was selected [4] .

construct modeling semantic space
Using morphological analysis to influence people perceptual image judgment modelling factor decomposition, looking for the individual factors of the decomposed variable element, as shown in figure 4, which will be the mascot of the shape of the deconstructed ten projects, 34 categories, including deconstruction projects form have head, mouth, eyes, nose, limbs and trunk, position can be divided into the extremities accounted for the proportion of the body, the rim of the eye of facial proportions, the transition of the neck, overall relationship has a head body scale, etc. The morphological elements of the mascot are respectively represented by a-h, and the characteristics of each morphological element are respectively represented by serial Numbers [5] , as shown in table 3.

establish a correlation model between perceptual images and mascot modeling elements
According to the structural results of the constituent elements of the mascot design, and combined it with the independent variable theory of science I theory, the division results of 21 representative samples under all the modeling features can be calculated. Thus, the deconstructed modeling elements are taken as projects, the modeling design elements are taken as categories, and on the basis of re-coding the modeling elements, in order to further quantify the modeling elements of the mascot samples, The matrix thinking can be used to quantitatively transform it into an editable data format. The specific method is: through the analysis of the characteristics of the modeling elements of the mascot, the independent variable corresponding to the characteristics of the corresponding elements is assigned to 1. Other attributes that do not match are assigned to 0 [6] . Finally, after analysis, we can only get the matrix containing "0" and "1", as shown in Table 4. y  y = g a1 a 1 +g a2 a 2 +g a3 a 3 +g b1 b 1 +g b2 b 2 +g b3 b 3 +g c1 c 1 +g c2 c 2 +g c3 c 3 +g d1 d 1 +g d2 d 2 +g d3 d 3 +g e1 e 1 +g e2 e 2 +g e3 e 3 +g e4 e 4 +g f1 f 1 +g f2 f 2 +g f3 f 3 +g g1 g 1 +g g2 g 2 +g g3 g 3 +g g4 g 4 +g h1 h 1 +g h2 h 2 +g h3 h 3 +g i1 i 1 +g i2 i 2 +g i3 i 3 + m y --Average value of perceptual semantic evaluation; g ij --Weight coefficient of each independent variable; a i～ i i --Modeling element response value; m --Constant term. Through statistical software SPASS linear regression analysis, which needs the input of the dependent variable is q is obtained by a questionnaire test on average, the evaluation of the emotional vocabulary and need to participate in the linear regression analysis is mascot representative sample modelling characteristic element matrix quantitative values, thick model parameters related data, including the partial correlation coefficient, category score, multiple correlation coefficient, constant value, such as in table 5. Therefore, according to the analysis results shown in Table 5, we can see that the "individual-ordinary" perceptual image is related to the modeling elements of the mascot.
The associated multivariate linear prediction model is: Y =0.041a 1 -0.112a 2 +0.197a 3 +0.835b 1 +0.547b 2 +0.306b 3 +0.406c 1 -0.173c 2 +0.634c 3 -0.255d 1 -0.117d 2 +0.671d 3 +0.213e 1 +0.503e 2 +0.2e 3 +0.082e 4 +0.463f 1 +0.244f 2 +0.380f 3 -0.484g 1 -0.345g 2 +0.14g 3 -0.276g 4 +0.217h 1 -0.121h 2 +0.112h 3 -0.003i 1 +0.521i 2 +0.371i 3 +0.345 Figure 6. Score of each modeling element Complex correlation coefficient is an index that reflects the degree of correlation between independent variables and dependent variables of a model. the larger the coefficient data is, the higher the linear correlation degree between elements and variables is. In order to verify the accuracy of the model, the square of the complex correlation coefficient is usually used to judge the determination coefficient. In a prediction equation, the closer R is to 1, the higher the accuracy of the prediction model. Generally speaking, the determination coefficient above 0.8 means that the fitting accuracy of the equation is better [7] . It can be seen from figure 6 that the R values representing the relationship between the morphological elements and perceptual images of the sample are all greater than 0.7, indicating that there is a very strong relationship between the observed value and the predicted knowledge [8] 。

conclusion analysis
The magnitude of the partial correlation coefficient reflects the influence of each project on the perceptual image of the mascot. As for the perceptual image phrase "individual-ordinary" in Table 5, it is analyzed from small to large according to the partial correlation coefficient shown in the table. it is concluded that the influence degree of each item is as follows: the ratio of head to body &gt; the transition of neck &gt; the ratio of eye circle to face &gt; the ratio of leg length to body &gt; the shape of mouth &gt; the shape of nose &gt; the ratio of hand length to body. By obtaining the degree of influence of the mascot element attributes on the emotional image, it can be used as a reference for the design process, that is, when the relevant emotional feelings need to be expressed, priority should be given to the influential mascot attributes.
The category score indicates the influence degree and emphasis direction of different categories on perceptual images in each project, the positive value indicates the positive meaning, and the negative value means the opposite meaning. Taking personality as an example, as shown in figure 6, we can see that the mascot modeling elements category in the case of "rounded rectangular body b 1 " has the highest contribution to it, followed by "heart-shaped nose d 3 "; in addition, the shape of the torso is "oval b 2 ", the neck transition "arc trapezoid i 2 ", the mouth shape is "arc e 2 ", "and so on also positively affect the" personality "image of the mascot. The negative part of the category shows that "nose shape" and "the proportion of hand length to body" have a high degree of influence on the "ordinary" image of the mascot, and the "common sense" is the strongest in "round triangle D 2 " and "1:2g 2 ". Therefore, in the design of mascots that need to reflect high-level feelings, because it is most closely related to ", the first thing that needs to be considered is that the topic is based on harmony, and it is the second item that needs to be paid attention to.

Conclusion
On the basis of Kansei engineering, this paper makes a quantitative study on the emotional tendency of the mascot to the audience. The semantic difference method is used to obtain the audience's perceptual image evaluation of the mascot shape, and the factor analysis is used to extract the main factors. At the same time, the morphological analysis method is used to deconstruct the basic structural attributes of the font and divide the features. The quantitative I theory establishes a mathematical model between mascot modeling features and emotional tendency, which makes it possible to predict the emotional tendency of mascot expression according to the objective factors of mascot design. Thus, to a certain extent, it can provide guidance for the design of the mascot, that is, on the basis of determining what kind of emotion the designer needs to express, the designer should follow the characteristics of the mascot structure which has obvious influence on this emotion.