Visual Measurement Technology for Tactile Kansei Information
Robin Huang, Akira Harada, Yutaka Shimizu

Summary
The fact, that each Kansei of human can be influenced by the other senses in a complex process. Kansei-engineering has played an important role in the field of design in the past decades. Unfortunately, the question of how to measure Kansei has remained unanswered. Based on "intermodality" , this paper aims to develop a new method for measuring tactile Kansei information on the sense of touch by using a haptic method.

At the present stage, the output and input parts of C&C product have not been able to posite together. It means, users must watch the output device and operate without being able to see the input device. Research that seeks to applicate on User-Interface between sense of touch and visual acuity is high recommended.

1. Introduction
1. Background

Beginning in 1986, the word, "Kansei-Engineering", was used by Mr. K. Yamamoto (chairman of the MAZDA-Auto Group) to describe the philosophy of car-making. To date, the concept of "Kansei-Engineering" has played an important role in the field of Industrial Design as well as other sciences during the past decades.

"Kansei Engineering" has been highly evaluated as the advanced technology of product development, which expresses the consumer demand in the design domain. In order to create a living environment, in which people are able to attain a better comfortable lifestyle, "Kansei Engineering" has been functioned as an ergonomic consumer oriental technology on new product development. Based on the human cognitive model, "Kansei Engineering" has attempted to create a pleasing and satisfying product for each user and consumer.

That is, "Kansei Engineering" can be defined as translating the technology of consumer's “Kansei" (physiological sensibility and psychological feelings) on new products into the design domain. Unfortunately, the question of how to measure the Kansei of a human being has remained unanswered.

2. Intermodality

Each modality of senses is not independent, but interacts with the others in a complex process. First of all, researchers of experimental psychology have attempted to establish several kinds of interactions among sensory modalities.

Previous psychology researches on sensation and perception have shown that the "Intermodality" phenomenon is existing. The condition of intermodality, in which a stimulus receives one sensory modality and reacts in another sensory modality subliminally. For example, the people somewhat can "hear" (feel) the bright or dark sound without seeing a thing. This phenomenon shows clearly that there are some innate qualities acrossing the different sensory modalities.

Secondly, each sensory acuity of human beings can be influenced by the other senses, which, in turn, have an effect on the advanced components of actions. This effect needs to be taken into account when it comes to the measurement of sensitivity. For example, even a slight action can induce interactive functioning of more than one sense at the same time, which is the phenomenon of "intersensory facilitation & inhibition".

Finally, the unique condition of "synesthesia", in which a stimulus receives one sensory modality and gives rise to an experience in another, has been known to the scientific community for nearly 300 years and yet remained relatively unexplored.

This aim of this paper is to develop a new method for measuring Kansei based on these phenomena.

3. Measurement

The basic theory and the model of the quantificational measurement is shown in Figure 1. The stimulus is an input ( = y ), and the response is an output ( = g ), human’s Kansei ( = x ) produced in the mental process is like a black box, because it is difficult to obtain y and g . We also can approach the Kansei x by another way. Suppose there is a dummy value x, and there is a function ( = f ) between x and x . Therefore, we can measure clearly the input y to be valued as h and the output g to be valued as g. Finally, we can calculate the value of x, and then approach the x .

Fig.1 Model of quantificational measurement

Basically, previous investigators have used the "expression law" and the "impression law" to assess the scale of Kansei of the sensory modality.

The expression law, usually be used in psychophysics and experimental psychology, measuring the sensibility by special equipment and obtaining a quantificational value-like data. The measurement like this may be understood as the response of physical sensibility, although it cannot describe the mental feeling of human’s Kansei completely. Nonetheless, the extension law, which measures the sensitivity through the physiology-like data, cannot be conducted perfectly. The method by the expression law has been tried to approach the value of g .

On the other side, the impression law, which tries to assess the data on the mentality information through a precisely fixed quantity scale, is still under development. Social science scholars, especially psychologists and cognitive science researchers, like to use the measuring method as this way to obtain the sensory attributes of people. They tend to ask testee directly by using the questionnaire to get the responses. The famous "SD method" (Semantic Difference method) is a typical method to reach the numerical data. Obviously, the method by the impression law has tried to approach the value of x.

4. Purpose

Being translating the technology of psychological feeling into the design specifications, we first need to develop a complete measurement of Kansei for the "Kansei Engineering".

For the sake of developing a useful measurement to evaluate the Kansei of human beings, this study proposes that it is possible to use the phenomenon of intermodality to measure quantificational Kansei by cross-modality matching as well as by the traditional methods of questionnaire.

So far, the research on tactile in psychology and the other files has been very limited. Therefore, as an example, we tried to measure the tactile Kansei by a new way to prove our hypothesis. Built from intermodality visual impression, the purpose of this exploratory study try is to shed new light on the measurement of tactile sensitivity and feeling by using a new haptical method.

Form the standpoint of Kansei Engineering, we have measured human’s mental Kansei based on the impression law and developed this new procedure to assess the tactile acuity of coarse.

2. Method and Experiment
1. Material
1. Tactile stimulus objects:

Four levels of different coarse sandpaper, as the tactile stimulus of coarse objects, were used in this experiment. These are #60, #120, #240, and #400 of the JIS standard. For the purpose of comparing 30mm×30mm with 200mm×200mm, two kind of shapes were used.

2. Visual expression tool:

Subjects were asked to adjust the visual noise effect of Adobe PhotoShop 4.0 on the monitor of the computer. The range of noise amount can be adjusted from amount 1 to 999 (see Fig 2).

Fig.2 adjustment of visual noise in this dialogue

3. Environments condition

To avoid interference with the effect of intersensory facilitation & inhibition, all experiments were constructed to stay in the same environment as possible. We used the same monitor and the same setting to avoid different lighting situations, control the noise condition of testing environment in the same situation, and record the temperature of each time (distributed from 22.3 to 27.0℃).

2. Method:
1. Subjects

20 subjects (10 male and 10 female), from 20 to 33 years old, were tested in this experiment. Every one of subjects was measured his/her body temperature before the experiment (distributed from 36.28 to 36.82℃).

2. Situation
   Each subject went through each of the following experimental situations as follow, and there were 4 levels of coarse stimulus in each experimental situation:

3. Procedures
   All subjects were asked to touch a coarse object (sandpaper) as the tactile stimulus without being able to see the object, and to adjust the noise effect of a similar-size gray area on the screen of monitor when they were touching the object. Subjects then were presented their physical stimulation by a way of " visual coarse " (adjust the noise effect of a gray square area). (see Fig. 3)

Fig.3 Subject be tested in this experiment

To avoid the sensory inhibition phenomenon, each experimental situation was held in a short period of time, with breaks after per 15 minutes. Time spent for each subject was recorded (the maximum time is 57 seconds, and the minimum is just 3 seconds).

3. Results
   The adjustment of the scale of visual noise-effect was recorded for each subject. 20 subjects, stimuli of 4 levels, were asked for 8 times of experiments in different situation. Thus, 640 results of response were obtained. Partial of the results was summarized in Table 1.

1. Observation

This experiment found an interesting pattern of the stimulus and the response. A mathematical relationship appears to exist between the tactile coarse stimulus and visual noise response.

Because the ratio of the original stimulus to the tactile coarse is 60: 120: 240: 400, it was summarized as 1:2:4:6.6 for purpose of calculation. As expected from the standpoint of Kansei engineering, the scale of Kansei is difference for each person. For visual observation, transformation the data into graph chart was made for subject (see Fig. 4). Hence, a nonlinear relationship in high dimension was observed from this distribution map.

Fig.4 response for different stimulus (subject No.01)

2. Trend analysis

For checking the correlation between the independent variable (i.e., value of tactile coarse stimulus) and the dependent variable (i.e., amount of visual noise effect), the "Trend Analysis" was used to analyze the data for each subject.

First, the general factorial ANOVA was conducted to test the P-Value is under 0.05 or not. Then, the Trend Analysis of MANOVA implemented in SPSS was carried out. These results were summarized in Table 2.

3. Correlation analysis

To determine the correlation between the tactile coarse stimulus and the visual noise effect, the "Correlation Analysis" was conducted on each subject.

Except for four subjects with a Pearson's correlation coefficient greater than 0.840, all other subjects were significant at the 0.01 level with a 2-tailed test (Tab. 3 and 4). In addition, the graphs of "Correlation Analysis" seem to indicate a nonlinear relationship among the subjects.

Table 3 Result of Correlation analysis (subject 1 and 2)

Table 4 Result of Correlation analysis (all subjects)

4. Regression analysis
   Finally, the "Nonlinear Regression Analysis" was used to estimate the association between the tactile coarse stimulus and the visual noise effect by each subject’s data. Different models of "curve estimate" (nonlinear estimate) were examined and tested for each subject (see Tab.5). The best fitted model for the "Cubic Pattern" ( Y = b0 + b1t + b2t ²+ b3 t ³) or "Growth Pattern" ( Y = e(b0+b1t)) or "Exponential Pattern" ( Y = b0*e(b1t)) were identified in mostly subjects.( "t" means the tactile stimulus, "Y" is the response in there.)

   Table 5 Result of Regression Analysis (subject No.01)

   Thus, used the "Nonlinear Regression Analysis" again to examine for the total average value of all subjects. And three formula as follows can be found obviously.(see Tab.6 and Fig.5) All of the R² value (Rsq.) of these three curve estimation are above 0.97.

   Y = (1.7E-05) t ³ – 0.0068t ²+ 1.0061t – 23.064

   Y = e(2.1925 + 0.0088t)

   Y = 8.9578*e(0.0088t)

Table 6 Result of Regression Analysis (total average)

Fig.5 Curve Estimation by (total average)

4. Implications:
1. Discussion
1. Relation across modality

The results of this experiment indicate a relationship between the tactile coarse stimulus and the visual noise response, and the tactile coarse reception is more sensitive than visual expression scale. In addition, this relationship seems to be consistent with a nonlinear correlation. For the author just used 4-difference tactile coarse stimulus in this experiment, so it’s difficult to judge the cubic relationship (Y = (1.7E-05) t ³ – 0.0068t ²+ 1.0061t – 23.064) is absolute or not with certainty.

In the other side, through this experiment, there are two relationship result, Y = e(2.1925 + 0.0088t) and Y = 8.9578*e(0.0088t), were identified also. Based on the famous equation of "Weber-Fechner’s law" in psychophysics, we know that if the E is the value of sensitivity and the R is the value of stimulus, then there is a relationship. Accordingly, we can suspect that this law is fit across two different modalities.

2. Cross-modality matching
   By comparing the result among subjects, we find the strong relationship between the dependent variable (amount of visual noise effect) and the independent variable (stimulus of tactile coarse).

   Although the gap between two modalities is existing, there is somehow an unification among different Kansei in the intermodality phenomenon. Thus, we may be able to estimate the Kansei-Value by cross-modality matching to obtain the interval scaled data.

   Of course, Kansei, as a complex combination of feeling and perception, is not easy to measure it clearly and accurately. Compared with the traditional questionnaire, measurement by cross-modality matching can obtain the metric data easily.

2. Conclusion
1. Theoretical model

This paper describes the qualitative evaluation of coarse tactile Kansei by using the cross-modality matching with visual noise impression. The author tried to build a theoretical model to explain how this new measurement can be used to approach quantificational Kansei.

For example, generally when a person receiving a coarse tactile stimulus (st) from the outside, he/she has used all sensibilities to perception factually in the nature situation. It means, he/she also receives the visual stimulus (sv) and the others kinds of input in the same time. By each modality, this person receipt/fell the different perception. In tactile modality, he/she fell the tactile perception xv(see Fig.6), and also he/she feel the visual perception xv in visual modality (see Fig.7), too. Than, he/she matches the perceptions cross the different modalities and integrates to a Kansei to output. Thus, we can build a model like Fig.8 showed.

If this model is tenable, it will be helpful for designer to develop the user’s interface in future. Because the coarse tactile reception is more sensitive than visual noise expression, therefore the designer maybe can design the buttons and switches by different coarse surface, as a useful hint to guide users for using the products.

Fig.6 Model of tactile modality

Fig.7 Model of visual modality

Fig.8

　

2. Be a new measurement technology
   The matching across the different modalities sums up the Kansei. Thus, it’s possible to obtain the value of x t (= tactile Kansei) from the amount of xv(= visual perception) , if there is some fixed relationship between xv–value (tactile perception) and xv–value (visual perception). We can questionnaire the subjects by visual noise expression to measure the coarse tactile Kansei.

   This type of study has important implications. For instance, visual image data can be used to define and classify the sensitivity of touch. Further, the obtained Kansei information on touch can be organized in a way to reflect the visual image of sense of tactile. Thus, this study helps to clarify the issue that sense of tactile and visual acuity might reflect "unification of a sense."

3. Limitation and Future Researches:
   What is most notable is that the author tested the possibility from the tactile coarse stimulus to the visual noise response in this experiment, difficult to say the cross-modality matching method is a completely useful measurement technology with certainty. In the future studies, it is necessary to test whether others tactile stimuli and/or others sensory modalities can also be express in visual modality. These are also the issues that the authors hope to make contributions. Finally, at this starting point for the development Kansei Engineering, more further studies are needed to develop a better measurement for Kansei.