Keywords

1 Introduction

By product ordering duties at retail store like convenience store, a salesclerk with ordering authority orders while watching an order terminals. In case that kinds of products are foods, if a product image is an image feeling that a salesclerk would like to eat it, he/she would like to order this product positively. To do it like that, what should provider of photographs do for them? [1,2,3] There are many factors from the view point of information design, they are, layout of screen interface, character color and font, so on. As a part of fundamental research about information design of sizzle image, representation of appetite by food image are studied.

Appetite like sizzler is defined as “advertising photograph of food or drink which stimulates feel of desire for food or drink. Sense which appeal their daintiness and freshness to customers” by Japanese dictionary [4]. To consider appetite by food image displayed on an order terminal, the effect to which salesclerk’s order is encouraged could expect.

The objective of this research is to extract evaluation items to assess appetite by food product image. To achieve this objective, two experiments were carried out. Experiment 1 tried to extract word sample to represent difference between images which feels appetite and images which not feel appetite. By using the samples extracted from this experiment, a list of evaluation items were developed by grouping same meaning words according to the descriptive evaluation method defined by JIS-Z9080 [5]. Experiment 2 was carried out by using these evaluation items and extracted items that evaluation results were different between two kinds of images, they were images which feels appetite and images which not feel appetite.

2 Experiment 1

Totally 33 graduate and undergraduate students were participated in this experiment. They are all normal and corrected normal vision.

22 images of food and drink which sold in general convenience store were used as stimulus images to show. Their breakdown were eleven kinds of food advertising pictures and food catalog pictures, respectively. The former defined images which feels appetite, the latter defined images which not feel appetite. Eleven kinds of foods were “Cheese hamburger”, “Chinese rice bowl”, “Iced coffee”, “Meat doria”, “Shrimp doria”, “Japanese dumplings”, “Spaghetti with meat sauce”, “Melon bread”, “Spaghetti with bacon and mushroom”, “Cold noodle” and “Salad”.

This experiment was carried out in the laboratory without windows. The environment condition such as illuminance in the laboratory was kept constantly. The participants evaluated independently. Choose one of eleven kinds of foods, both food advertising picture and catalog picture were shown to a participants. They were required to describe words that they hit on to sticky about the difference correspond catalog pictures to advertising picture. Time limit does not set. The experiment closed that the participants judged that they took out words up. During the experiment, experimenter waited outside of the experiment room.

3 Grouping Words

From the results, 350 words could be collected. The objective of this section is to make evaluation items to be able to assess appetite using word samples collected in the experiment 1. To achieve this, grouping used collected word samples was carried out according to the descriptive test method in JIS-Z9080 without considering difference of foods and gender of observers.

3.1 Method

The task of grouping was carried out through discussing six experimenters related to this research. Using sticky with evaluation words, grouping was performed every word which means the same contents.

3.2 Result

Finally, 37 groups were made by using 280 words shown in Table 1. During the discussion, the words do not related to appetite were not used in the analysis. They are about the way to take a photo (e.g. position of a camera, light condition, beauty) and so on.

Table 1. Result of grouping by same meanings

4 Experiment 2

The objective of this experiment is to verify that each item of evaluation word list made in Sect. 3 is suite for evaluating appetite by food image. To achieve this, evaluation items which are able to evaluate the difference between two images (they are images which feels appetite and images which not feel appetite) were tried to extract by evaluating these images using each item.

4.1 Experimental Method

Ten graduate and under graduate students who did not join the experiment 1 participated in this experiment. They are all normal and corrected normal vision. 33 evaluation items listed on Table 1 which were extracted based on 37 items obtained by Sect. 3 were used. Procedure of extraction is shown below.

Firstly, “yearning”, “calm” and “elegant” were deleted from the evaluation items because the number of samples of each item is only one, respectively. Remained 34 items were classified into four groups, that is “sense”, “first impression”, “estimation by vision” and “beautiful”, and “looks tasty”, “satisfaction” and “friendly” were classified into “delicious”, “satisfaction” and “friendly” were classified into total evaluation. Lastly, about “good scent” and “grilled flavor” classified in “estimation by vision”, the latter was deleted because meaning of these two words were similar. From these procedure, 33 items were listed shown in Table 2.

Table 2. Evaluation items using the experiment 2

The experimental environment was same as experiment 1. The evaluation item in Table 2 was lined up in turn, five step scale was located on the side of each item [6]. Evaluations were carried out by participant’s own pace without time limit. The participant was required not to think too deeply and to judge by the intuition as much as possible. During the experiment, experimenter waited outside of the experiment room.

4.2 Results and Discussions

From the results of evaluation data by using 33 items, average and standard deviations of evaluation data from 10 participants by each food image were calculated. Parts of results about “sense” were shown in Figs. 1 and 2, parts of results about “estimation by vision” were shown in Figs. 3, 4 and 5, and parts of results about “total evaluation” were shown in Figs. 6 and 7.

Fig. 1.
figure 1

Mean evaluation value of “warm”

Fig. 2.
figure 2

Mean evaluation value of “cool”

Fig. 3.
figure 3

Mean evaluation value of “newly made”

Fig. 4.
figure 4

Mean evaluation value of “juicy”

Fig. 5.
figure 5

Mean evaluation value of “healthy”

Fig. 6.
figure 6

Mean evaluation value of “satisfaction”

Fig. 7.
figure 7

Mean evaluation value of “friendly”

To extract items which shows the difference between images feel appetite (food advertising picture) and images not feel appetite (catalog pictures), t-test using average data between two conditions were carried out. From the results, two items (“warm”, “bright”) have significant difference, they were, (t(20) = 2.25, p < .05) and (t(20) = 4.12, p < .05), respectively.

There is no significant difference in all items in “cool” (t(20) = 1.67, ns). In “first impression”, except “invigorating” (t(20) = 1.08, ns) and “young and vivacious” (t(20) = 1.08, ns), there are all significant difference. About “estimation by vision”, nine items have significant difference (e.g. “newly made (t(20) = 5.37, p < .05), “hand made” (t(20) = 4.87, p < .05)).

About “total evaluation”, all three items have significant difference (“looks tasty” (t(20) = 6.06, p < .05), “satisfaction” (t(20) = 4.38, p < .05), “friendly” (t(20) = 3.67, p < .05))

24 items which have significant difference between conditions got higher evaluation value for images feel appetite (food advertising picture) than images not feel appetite (catalog pictures). From this, these 24 evaluation items could evaluate “feel appetite”.

Osgood [7, 8] described that there were three universal axes for evaluating things, they were “activity”, “evaluation” and “potency”. The rightest column in Table 3. Shows the results of applying these three words to evaluation items. From this, almost words were classified to “activity” or “evaluation”. There was few “potency” because the objects of this time were related to food. So, the evaluation items obtained this research was validate from the view point of evaluating things.

Table 3. Mean evaluation value of “friendly”

5 Conclusion

From the results obtained through the experiments, evaluation items for “feel appetite” could be narrowed down 24 items shown in Table 3. By using these items for evaluating food image, “feel appetite” is expected to evaluate.

To generalize these scale in the future, it is necessary to assess validity of these 24 items. As a concrete procedure, it is considered to verify the relationship between appetite by food image and score of 24 evaluation items through evaluation experiment using various food image not using this time experiments and to clarify the structure of “feel appetite” factors by extracting common factors between evaluation items using statistical methods. As there is a research to develop a method for representing “feel appetite” physically [3], quantification the relationship between physical value and psychological value of “feel appetite” could be expected.

This time, photography experience of participants and view point of photography were unknown. These factors are efficient for evaluation results, and angle of photography and background design are also efficient for evaluation results. So, experimental condition should be clarified.

To the future, an evaluation method of “feel appetite” could be established for verifying validity of evaluation items by trying additional experiment or additional analysis.