Keywords

1 Introduction

The period of the Republic of China was important for the transformation of Chinese advertising from tradition to modern. It was not only the beginning of our modern advertising, but also the first development peak of our advertising. Advertising can be seen as the living fossil of the history. Through the study of advertising during the period of the Republic of China, our historical and cultural heritage, from different dimensions, we can understand the width and height the design practice can reach and feel the economic development status of the society in that period which is also called as the early stage of modern design in China.

Culture heritage is the essence extracted from the people’s work and living after a long time, and the fusion of a variety of traditional life inherited by generations and closely related to the life of people. With society gradually paying attention to the protection of historical and cultural heritage, more and more libraries and museums concerning the intangible cultural heritage are built. The expression of culture heritage should combine the carrier itself with experience and feelings of people during visiting to make visitors not only have certain knowledge of the carrier but also get a quiet different interactive experience including emotional and cultural awareness and aesthetic, etc. through the interactive relationship between visitors and museum.

In order to abstract design elements which meet the demands of contemporary aesthetic and are in accord with the experience of modern aesthetic, we choose 571 pieces of print ads during the period of the Republic of China as samples to analyze their style and characteristics based on the requirement for culture heritage protection.

2 Design the Research

This paper will do research from the following aspects:

  • Using a custom-developed software to identify the image feature, abstract the information of image feature significantly related to advertising sample and form the variables indicators.

  • Basing on the technology of feature analysis, study the specific attributes in information architecture, visual representation for information, the aesthetic characteristics and construction method, technology and cultural elements, etc. and form the research variable indicators in the process of advertising samples labeling.

  • Explore the image significant correlation between the objective elements and subjective cognitive elements and find the main features by analyzing the research variable indicators with statistical calculation method and technologies, such as cluster analysis, factor analysis, correspondence analysis, multidimensional scaling analysis, etc.

  • Provide professional and effective guidance for the design of museum basing on the research result, including the overall atmosphere, thematic construction, situation creation, etc.

The samples used in this paper were popular with the economic zone taking Shanghai as the focus during the period of 1910-1940. Figure 1 shows that it contains pieces of print ads, such as posters, paper wrappings, cards, etc. In order to do research, the following pictures are all coded.

Fig. 1.
figure 1

571 pieces of print ads (partial)

Full size image

3 Research

3.1 Procedures

  • Research on recognizing the image feature of advertising samples basing on calculation: Through chromatographic, color tune and singularity analysis, find the image feature with which sample variable indicators have a tight correlation, extract and label it.

  • Image analysis, variable definition and labeling basing on feature analysis (do parallel with step 1): do feature analysis for all the visual advertising samples, make a label list and form nominal level, ordinal level and scale variables for the subsequent statistical analysis.

  • Data mining basing on multidimensional indexes: basing on the data achieved from step 1 and 2, excavate the dominant or potential significant correlation in multiple dimensions and multi-facts, build the framework of influencing factors and extract the common and individual feature of samples.

3.2 Research on Recognizing the Image Feature of Advertising Samples Basing on Calculation

The image feature of advertising samples contains several different dimensions. In this paper, image feature refers to image color feature which includes theme tune and representative color.

In this paper, all samples are JPG format data files and their color models are RGB. In the past, some color reduction research for the monuments establishing color reduction algorithm with the help of implements’ real color. It made the precedent in color reduction for the monuments. But in this paper, the real objects can’t be found to apply the reduction algorithm because the real objects in the picture of samples are abundant, and the samples span 30 years. Because of the limit and in order to protect the initial data, this paper will analyze directly the samples by RGB color model.

The research tools for recognizing the image feature in this paper are developed by our research team with JAVA language, including color spatial distribution analysis tool, theme tune extraction tool and theme tune similarity analysis tool.

Color spatial distribution analysis tool. Color spatial distribution analysis tool generates color points according to pixels in samples (Fig. 2), forming a color space, clusters the points and reduces their dimension by vector computation and principal component analysis to help researchers analyze the color feature of samples.

Fig. 2.
figure 2

Color spatial distribution analysis tool colorSpace_3D_v1

Full size image

Theme tune extraction tool.

Theme tune extraction tool does color feature calculation of pixel, obtains the constituent of global color tune separately (Fig. 3), then extracts component of the theme tune which can describe 80 % tune in samples (Fig. 4), does batch processing and analysis and generates a theme tune table of every sample for the following multidimensional scaling analysis among samples.

Fig. 3.
figure 3

ImageColor_Abstraction_and_Sorting_v2

Full size image
Fig. 4.
figure 4

ImageColor_Abstraction_and_Sorting_v3

Full size image

Theme tune similarity analysis tool.

Theme tune similarity analysis tool (Fig. 5) calculates the similarity of the theme tune between any two and generates a table with.csv format. The table will be used for the following multidimensional scaling analysis and statistical calculation of MDS.

Fig. 5.
figure 5

The output result of Themetune similarity analysis tool

Full size image

3.3 Image Analysis, Variable Definition and Labeling Basing on Feature Analysis

The main task of this paper is to form a structural system of variables to describe samples. After labeling the system, form a normal and ordinal description for variables to support the following descriptive statistics and correspondence analysis and excavate the common and different feature among samples.

This paper raises several variables for description basing on the overall analysis of the 571 pairs of advertising and labels the variables. The variables are classified into 6 categories.

  • Content label (theme, product, product type, place of production, consumer orientation, brand, etc.);

  • Composition label (diagonal, hub-and-spoke, zigzag, revolving shape, etc.);

  • Expression label(direct display, emphasize on feature, contrast and foil, reasonably exaggerate, see big things through small, associative thinking, rich in humor, metaphorical transfer, use wonderful feeling to foil the product, suspense, idol, imitation, mystical method, multi pictures, etc.);

  • Visual expression label (oil painting style, printing style, watercolor style, photo style, line drawing style, ink style, meticulous style, etc.);

  • Image theme label (person, animal, objects, scene, etc.);

  • cultural label (the traditional oriental, the west);

3.4 Measure Data Matrix

The data matrix is a contingence table used for management of samples, the corresponding labels and value. All the value is combined by image feature and artificial label. After the necessary standardization, all the data will be Imported into the SPSS software to do the following descriptive statistics and advanced statistical calculation (Table 1).

Table 1. Sample measure data matrix sheet
Full size table

3.5 The Statistical Calculation and Analysis Basing on Multidimensional Indexes

In this paper, qualitative research and quantitative research are the main method to analyze the information and data. The data required from the process of feature analysis are mainly nominal data, for example, the expression methods used in the sample. It offers data for analyzing the variable distribution in all samples.

Owing to the large number of variables and there are more specific variables and value in six categories, this paper adapts correspondence analysis and multidimensional scaling analysis to dig the feature of samples basing on basic descriptive statistics.

4 Analysis

4.1 Result Analysis of Recognizing the Image Feature

The following are the results basing on the similar color space positioning analysis and calculation of samples (Figs. 6, 7 and 8).

Fig. 6.
figure 6

Position of the themetune of all samples

Full size image
Fig. 7.
figure 7

Position of the themetune rotated of all samples

Full size image
Fig. 8.
figure 8

Position of the themetune of all samples whose dimension is reduced

Full size image

Figure 8 shows that it is clearly seen that the position of all samples in two dimensional spaces are in accordance with the rule of two dimensions:

  • In the horizontal dimension, the theme tune shows the transition from black and white color (left) to high saturation color (right).

  • In the vertical dimension, samples show the transition from cool tone (down) to warm tone (up).

  • The distribution of samples in the horizontal direction is wider than in the vertical direction, i.e. saturation is more representative than color temperature for the difference of the theme tune.

  • The number of samples gathering in the center of the picture is bigger than that in the surrounding area, i.e. the theme tune of the most samples are similar. It reflects the relative uniformity of the theme tune.

The following is the result of multidimensional scaling analysis of MDS using similarity analysis tool of theme tune (Figs. 9 and 10).

Fig. 9.
figure 9

Position of tune similarity MDS

Full size image
Fig. 10.
figure 10

Position of tune similarity MDS (including color tags) (Color figure online)

Full size image

Figure 9 shows that it can be easily seen that similarity distribution of the theme tune of all samples shows significant rule. We can achieve the binary regression equation by using twice curve-fitting method.

$$ {\text{Y}} = 0. 3 3+ 0. 2 5*{\text{X}} + 1.0 9*{\text{X2}},{\text{ R2}} = 0. 5 2 7 $$

In order to make the analysis result easier to find the distribution rule, researchers give more information to Fig. 9, add color code of theme tune to every sample (Fig. 10).

As can be seen in the picture, although the attribute and content of all samples are different, the feature of theme tune shows obvious rule. Color brightness changes gradually from bright to dark from left to right, but in the vertical dimension, no clear rule can be found. In addition, saturation in center section enveloped by U-shaped curve is higher than that around.

4.2 Result Analysis of Separate Identification of Label Attributes

The following are analysis result of position picture from different variable dimension.

Analysis about tone richness and analysis from other variable dimensions

Analysis about composition way and tone richness.

Figure 11 shows that there is no clear relationship between composition way and tone richness in all samples and there is no clear rule that can explain the distribution of composition way in all samples. In addition, it is rare that repeatedly arranged patterning method, diagonal composition method and symmetrical composition method are widely used in the samples.

Fig. 11.
figure 11

Analysis about composition way and tone richness

Full size image

Analysis about expression way and tone richness (Fig. 12).

From the result, most of the expression ways can be easily seen in the 571 samples, but several ways, including text announcement method, suspense arrangement act, humorous imitation method, see big things through small ones, etc., are less used than other ways.

Fig. 12.
figure 12

Analysis about expression way and tone richness

Full size image

Analysis about visual expression way and tone richness (Fig. 13).

From the result, calligraphic style is relatively rare, so are photographic style and painting style, but the latter are used more than the former, the expression of the most samples are similar.

Fig. 13.
figure 13

Analysis about visual expression way and tone richness

Full size image

Analysis about culture label and tone richness (Fig. 14).

In this picture, there is no clear relationship between culture label and tone richness in all samples. In this dimension, it is unique that historical figures and folk customs are used. A lower proportion of appearance may be the reason.

Fig. 14.
figure 14

Analysis about culture label and tone richness

Full size image

Analysis from other dimensions

Overall, 571 samples show the following common characteristics:

  • Textile products are rich in the expression way and ‘scenario contrast’ and ‘imaginations’ are most used. Cigarettes, drugs and cosmetic products prefer ‘idols’ followed by use ‘wonderful feeling to foil the product’, ‘suspense arrangement’, ‘humorous imitation’. Festive supplies use ‘direct display’ more. Daily utensils use ‘contrast’ more.

  • Textile products are also rich in culture label and it use ‘love’, ‘nature’, ‘story’, etc. Stationery commodity prefers ‘history’. Apparel fabrics use ‘idols’ more. Daily utensils usually use ‘the west’. Tobacco and festival activities use ‘a combination of Chinese and Western elements’ widely.

  • In consumer orientation and choice for culture label, cultural groups prefer ‘history’ label. ‘A combination of Chinese and Western elements’ and ‘Stylish in business’ labels are more used in the female consumer group. Smokers use ‘idols’ more. Professionals use ‘story’ more.

5 Conclusion

The research results can offer an effective guidance to the design of Professional museum for print ads mentioned above during the period of the republic of China, including the overall atmosphere, thematic construction, and situation construction of the museum.

The samples dominant tone presents obvious regularity in visual cognition. On the scatter diagrams, the tone of the horizontal dimension is clear changed gradually from light to dark. There is no striking rule on the vertical dimensions. The overall museum atmosphere should be consistent with the samples tone, and keeping responding relation in light-shade & cold-warm.

The sample theme should reflect its own cultural label, such as by telling historical stories, reappearing tales of legendia. The characteristics of the time should be fully embodied by the help of the special idols from the Republic of China, fashionable product and arisen combination elements of Chinese and Western.

The construction of the Museum should pay attention to create story atmosphere, such as stories based on story background, story connection, suspense, and passion things, which makes the viewers be personally on the scene of the stories of the ads. It reflects a unique expression to the ads of Republic of China.

Further more, we can adapt this visual-based approach to design elements research in cultural heritage study, that could help us to evaluate visual products, and to evaluate the effect of visual cognition. This approach has a very high important significance in cultural archeology and design industry research as well as other fields.