Keywords

1 Introduction

Visual search is the cognitive process of searching for a specific target object under a certain background [1]. It is not only an important method of daily learning and obtaining outside information but also a significant part of various professional tasks, for example, map reading, the X-ray luggage inspection, medical image analysis, driving and web searching [2]. As the dominant status of mobile internet is strengthened gradually, the user communities of various mobile phone applications are expanding and users’ visual search efficiency has been an important subject for improving user experience in mobile phone interactive interface design.

2 Research Status

According to The 40th Statistical Report on Internet Development in China, during the first half of 2017, applications on business transaction keep growing rapidly on China Mobile Internet and the characteristics of upgraded consumption in the online shopping market are more conspicuous, while user preferences are gradually shifting to quality, intelligence and new product consumption. Studies by people such as Aurora showed that image navigation can deliver more information to users, and abundant information can enhance the attractiveness to users as well as promoting users’ search behaviors [3]. Lulu took the WeChat platform as the research object. By analyzing its interface layout, images and visual performance, she found that images are more intuitive and easier to understand, besides, rich icon interfaces tend to be more attractive to users [4]. From the perspectives of visual search and visual browsing regularities, Ting Liu and others analyzed the influence of mobile phone news apps’ graphic and textual layouts on users’ reading efficiency, and studied eye-tracking regularities of visual browsing and visual search under different layout schemes [5]. In conclusion, most of the existing researches analyze users’ visual behaviors from the perspective of interface characteristics, which tend to analyze the cognition results, but seldom study users’ behavior regularities and behavior mechanisms. Based on the similarity theory of visual search processing, taking the shopping APP interface as an experimental sample and through the eye-tracking experiment, this study mainly researches users’ visual search efficiency and mechanism in different types of interfaces.

3 Similarity Theory and Interaction Design

In interactive interface design, the efficiency of visual search is related to the characteristics of the interface. In cognitive psychology, the efficiency of visual search is closely related to the way of human cognitive processing. Duncan and Humphreys put forward the similarity theory of visual search in 1989. In this theory, visual search is divided into two processes: perceptual description and reaction judgment. It emphasizes that efficiency of search task is determined by the difference between target stimulus and interference item and the similarity among the interference items [6]. Based on the similarity theory of visual search, taking the efficiency of visual search in mobile phone interactive interface as the experimental sample, this study mainly discuss the characteristics of visual search efficiency in different types of interfaces and analyze the forming system (Fig. 1).

Fig. 1.
figure 1

Visual search process under the similarity theory

Full size image

In this study, we take shopping App interface as the example. During the process of interaction between the user and the mobile phone interface, the process of looking for a target product by a user can be regarded as a process of visual search and matching. We know from the similarity theory that visual search efficiency is closely related to target items and interference items, therefore, by comparing and analyzing the interface elements of the mainstream shopping Apps, this study sums up following 2 group of variables at first stage: whether the product pictures contain background information (including shooting background, accessories and text labels, without them is a single product picture); whether the layout is regular (regularity means the same picture arrangement shows repeatedly, unregular layout means no picture arrangements are the same). By designing and conducting eye-tracking experiment, this study collects users’ eye movement and behaviors’ data to observe users’ visual search efficiency. Combining with the existing work, this paper made the following assumption: visual search efficiency is higher when users browsing pages with pictures that has no background information and with regular layout, otherwise, visual search efficiency becomes lower.

4 App Interface Search Eye-Tracking Experiment

4.1 Tested Objects

The experiment recruited 10 undergraduates which include 5 males and 5 females. Their ages are among 22–26 and all of them have experience with using shopping apps. They all have normal or corrected vision, no one has color blindness or color weakness.

4.2 Experiment Devices

Experiment devices include Tobii Glasses eye-tracking instrument, desktops (with 19 in. 1920 * 1080 resolution screen) and Tobii Studio 3.2 eye-tracking data analysis software. Before the experiment starting, every subject is calibrated. Sample pictures are of the same size to the mobile phone screens and shown on the 19 in. screen. Subjects sit keeping their eyes level with the screen and the distance between their eyes to the screen is 40 cm, close to the distance between their eyes to their mobile phone screens.

The environment of mobile interface is different from computer interface. Limited by equipment conditions, the mobile phone interface cannot be evaluated directly in this experiment. But previous research shows: user performance tests prove that availability assessment conducted by the computer simulated mobile phone prototype interface is effective [7].

4.3 Test Design and Material

This test is designed for the factors of 2 pictures’ background information (with or without) × 2 pictures’ layout regularity (with or without), and test material is analog shopping interface. By combining the variables, it formed 4 different kinds of test interfaces: interface of pictures without background information but with regular layout, interface of pictures with background information and regular layout, interface of pictures without background information or regular layout, interface of pictures with background information but without regular layout.

4.4 Test Procedure

The test is carried out in a quiet room. Before the test starting, subjects carry out nine-point calibration and learning test, keep their heads motionless after calibration and conducted learning test, then start migration test. The learning process is as follows: present the targets for 2s—present the fixation point on the upper left for 2s—present the test interface with 1 pictures—click the mouse to finish the task when find the target; The migration test process is as below: present the targets for 2s—present the fixation point 2s—present a migration test interface—find the target, click the mouse to finish the task and turn to the next page (4 times)—finish 5 tasks—finish the test (Fig. 2).

Fig. 2.
figure 2

Example of experimental process

Full size image

4.5 Data Processing

The 10 groups of data are all effective. Eye movement data is recorded by eye-tracking instrument, behavior data is counted by stopwatch, they are used to record the search time and fixation counts for studying. Test data are analyzed and processed by spss 23.0.

5 Data Analysis

5.1 Time Spent for Searching Targets

Through the statistics of time spent on turning the test pages, we get the descriptive statistics (Table 1) and results of multifactor analysis of variance for the search time on the test interface (Table 2).

Table 1. Descriptive statistics of search time on the test interface
Full size table
Table 2. MANOVA for search time on the test interface
Full size table

From the multifactor analysis of variance, we drew the following conclusion: pictures’ background information has an obvious main effect (F = 31.837, p < 0.05), which indicates subjects’ search efficiency is higher in interface of pictures without background information than in interface of pictures with background information. Layout regularity also have an obvious main effect (F = 40.935, p < 0.05), which indicates subjects’ search efficiency is higher in interface with regular layout than in interface without regular layout. Picture background information and layout regularity have significant interactive effects (F = 11.837, p < 0.05). Influenced by both factors, the search time subjects spent on different types of interface is significantly different.

According to the descriptive statistics of search time, pages of which pictures without background information and have regular layout take a shorter average search time than of those pictures with background information and have unregular layout. The independent t-test result is F = 68.802, p < 0.05, which indicates that the original assumption is incorrect. Therefore, pages of which pictures without background information and have regular layout take an obviously shorter average search time than of those pictures with background information and have unregular layout.

5.2 Total Fixations

By extraction and analysis on eye movement data through Tobii Studio and spss23.0, we get the descriptive statistics (Table 3) and multifactor analysis of variance for total fixations on test interface (Table 4).

Table 3. Descriptive statistics for total fixations on test interface (/times)
Full size table
Table 4. MANOVA for total fixations on test interface
Full size table

From the multifactor analysis of variance, we draw the following conclusion: pictures’ background information has an obvious main effect (F = 29.297, p < 0.05), which indicates subjects’ fixations on interface of pictures without background information are less than on interface of pictures with background information. Layout regularity also have an obvious main effect (F = 56.658, p < 0.05), which indicates subjects’ fixations on interface with regular layout are less than on interface without regular layout. Pictures’ background information and layout regularity have significant interactive effects (F = 11.604, p < 0.05). Influenced by both factors, subjects’ fixations on different types of interface is significantly different.

According to the descriptive statistics, interfaces of which pictures with background information and have unregular layout get more average fixations than of those pictures without background information and have regular layout. The independent t-test result is F = 15.202, p < 0.05, which indicates that the original assumption is incorrect. Therefore, pages of which pictures with background information and have unregular layout obviously get more total fixations than of those pictures without background information and have regular layout.

6 Experiment Analysis

In this study, visual search efficiency of mobile phone interface under different variable conditions is studied through eye-tracking test analysis and based on experimental data analysis. Below conclusion is drawn: affected by the variables of whether the pictures contain background information and whether the layout has regularity, subjects’ visual search efficiency varies significantly. Therefore, discussions in this paper are based on the two variables.

6.1 Visual Search Time

According to the experimental data, subjects obviously spent less time in interfaces of pictures without background information and have regular layout than in those of pictures with background information and have unregular layout. In previous studies, people like Scharroo considered that visual search efficiency goes down as pictures’ complexity level rises [7]; Jing Li and other found from studies that the type, adjacency state and connection mode of shape coding are important factors for whether it can be fast captured. Above findings are all in conformity with the experimental results of this study. In view of similarity theory, with pictures’ background information as interference item, the content for subjects to compare in visual search process is increased. Besides, pictures appear at a random position in interfaces with unregular layout, which decreases the similarity among interference items. Therefore, it takes more search time in interfaces containing pictures with background information and have unregular layout. In interfaces with regular layout, pictures are in the same position and without background information, which enables subjects to seize the content of pictures and then compare and judge quickly. That is to say, it decreases the similarity between targets and interference items, so interfaces without background information and with regular layout takes less search time.

6.2 Counts of Fixations

According to the test results, fixations on interfaces containing pictures without background information and regular layout is obviously less than on interfaces containing pictures with background information and unregular layout. As pictures without background information are easy to focus and regular layout interfaces guide users’ vision path regularly, which help to improve visual search efficiency, thus interfaces containing pictures without background information and regular layout get less fixations. Background information in pictures increases the information of targets, which makes the increased fixations in pictures with background information. Besides, unregular layout interfaces have disordered and possibly repeated vision path, which makes the relatively more fixations in interfaces of pictures with background information and unregular layout.

6.3 Deficiencies

After analysis of the results and profound consideration in the later stage, we find below problems: control variables are not clearly classified, which may affect the test results. Instead of motion environment, subjects are in static state when simulating the mobile phone operation in front of the computer screen, which may affect the test results. During the process of data analysis, we only discussed the two extremes of variables’ combined effect and need further simple effect analysis on the specific relations between the factors of dual effects.

6.4 Conclusion

Similarity theory is one of the visual search and processing theories based on external characteristics guidance. This paper is based on the theory and studied the visual search efficiency and its mechanism in interfaces with different characteristics. According to the results, we draw the following conclusions:

  1. 1.

    In interactive interface design of mobile phone, the effect of pictures’ background information and layout regularity need to be considered. Users’ visual search efficiency and cognitive performance can be improved by simplifying pictures’ background information and adding regularity of pictures’ layout.

  2. 2.

    Visual search efficiency is relatively higher in interfaces where pictures are without background information. Pictures’ background information would decrease visual search efficiency. In interactive interface design of mobile phone, pictures with background information of different complexity are needed according to different designs.

  3. 3.

    Visual search efficiency is relatively higher in interfaces with regular layout and would decrease in interfaces with unregular layout. Thus in interactive interface design of mobile phone, we should focus on interfaces in regular layout to reducing users’ cognitive cost and improve cognitive performance.

  4. 4.

    As users’ visual search efficiency takes the dual effect of picture background information and layout regularity, we should set users’ visual search efficiency according to different interface functions in interactive interface design of mobile phones. To enrich users’ operation performance, we can guide different visual search efficiency by changing the complexity of pictures’ background information and the regularity level of layout.

7 Epilogue

Based on the similarity theory of visual search, this paper mainly studies users’ visual search efficiency in different types of mobile phone interactive interfaces, and draw the conclusion that the factors of picture background information and layout regularity would obviously affect visual search efficiency. It also discussed the matching mechanism under the effect of the two factors. However, this study only discussed the combined effect of the factors, and our next study should include detailed analysis in specific situations of the factors.