Research on Generating an Indoor Landmark Salience Model for Self-Location and Spatial Orientation from Eye-Tracking Data

Landmarks play an essential role in wayfinding and are closely related to cognitive processes. Eye-tracking data contain massive amounts of information that can be applied to discover the cognitive behaviors during wayfinding; however, little attention has been paid to applying such data to calculating landmark salience models. This study proposes a method for constructing an indoor landmark salience model based on eye-tracking data. First, eye-tracking data are taken to calculate landmark salience for self-location and spatial orientation tasks through partial least squares regression (PLSR). Then, indoor landmark salience attractiveness (visual, semantic and structural) is selected and trained by landmark salience based on the eye-tracking data. Lastly, the indoor landmark salience model is generated by landmark salience attractiveness. Recruiting 32 participants, we designed a laboratory eye-tracking experiment to construct and test the model. Finding 1 proves that our eye-tracking data-based modelling method is more accurate than current weighting methods. Finding 2 shows that significant differences in landmark salience occur between two tasks; thus, it is necessary to generate a landmark salience model for different tasks. Our results can contribute to providing indoor maps for different tasks.

Retrieving Landmark Salience Based on Wikipedia: An Integrated Ranking Model

Landmarks are important for assisting in wayfinding and navigation and for enriching user experience. Although many user-generated geotagged sources exist, landmark entities are still mostly retrieved from authoritative geographic sources. Wikipedia, the world’s largest free encyclopedia, stores geotagged information on many geospatial entities, including a very large and well-founded volume of landmark information. However, not all Wikipedia geotagged landmark entities can be considered valuable and instructive. This research introduces an integrated ranking model for mining landmarks from Wikipedia predicated on estimating and weighting their salience. Other than location, the model is based on the entries’ category and attributed data. Preliminary ranking is formulated on the basis of three spatial descriptors associated with landmark salience, namely permanence, visibility, and uniqueness. This ranking is integrated with a score derived from a set of numerical attributes that are associated with public interest in the Wikipedia page―including the number of redirects and the date of the latest edit. The methodology is comparatively evaluated for various areas in different cities. Results show that the developed integrated ranking model is robust in identifying landmark salience, paving the way for incorporation of Wikipedia’s content into navigation systems.

The Influence of Spatial Familiarity on Landmark Salience Sensibility Based on Eye Tracking

<p><strong>Abstract.</strong> Landmarks play an important role in navigation especially when people in the familiar and unfamiliar environment. These landmarks are usually used for expressing their spatial knowledge. This prior knowledge is related to the spatial familiarity that the spatial knowledge is acquired by individuals as a function of their experience in the environment (Gale et al., 1990). Individuals prefers to take into highly account their familiar route within the environment. Furthermore, the route descriptions include the differed types of landmarks in familiar/unfamiliar routes (Lovelace et al., 1999). The user’s level of spatial familiarity for people is a key factor of how they navigate in the environment (Savage et al., 2012). In human navigation or wayfinding, the dependency on familiar landmarks is the highest priority (Golledge, 2003).</p><p>Landmarks are more distinguishably salient and prominent than the another spatial features (Sorrows and Hirtle, 1999). Therefore, the landmark salience refers to how easily this prominent spatial feature can be regarded as a landmark (Raubal and Winter, 2002). The landmark salience mainly divided into visual, structural (important location), and cognitive (semantic). Visual salience refers to the visual characteristics of spatial features such as color contrast with surroundings. Some visual characteristics include facade areas, shapes, colors and other properties. Semantic salience refers to the spatial knowledge related features such as a cultural importance of building. Structural salience refers to the spatial features play an important role in the structure of the spatial environment.</p><p>The level of spatial knowledge of the individual also an important factor in landmark selection (Nuhn and Timpf, 2018). Hamburger and Röser (2014) investigated landmark semantics and showed that famous buildings were better recognized than unfamiliar ones. Quesnot and Roche (2015) found that the familiar individuals prefer the semantic landmarks in environment while the unfamiliar individuals prefer the landmark with the prominent visual and structural. In contrast, people unfamiliar with an environment prefer landmarks with outstanding visual and structural characteristics.</p><p>The existing literature studies the method that landmark indicators to be considered are not comprehensive while can’t reflect on the landmark characteristics. The reason is that the method of calculation and extraction for landmarks can’t conform the individual preference needs of navigation. Compared with the traditional subjective evaluation method, the eye-tracking method provide objectivity and reliability. For example, a mobile eye tracking system was used by Kiefer et al. (2014) to investigate the process of self-localization. They discovered that test persons directed their attention longer on landmarks on the map and aligned them with objects in the surroundings during successful localization of their own position. Schwarzkopf et al. (2013) used fixed and mobile eye tracking devices to study that participates selected the navigation landmarks in the virtual environment and the real-world of the airport. In general, there are an only limited number of works related to the influence of spatial familiarity on landmark by using eye tracking method.</p><p>At present, many navigation systems are providing users with route directions under different travel modes. These navigation systems have the "distance-to-turn" method in route directions, however the landmark knowledge has not been conformed to user's habits and spatial cognitive in wayfinding. In order to solve these problems, this paper aims to study the characteristics of the user’s spatial familiarity and landmark salience evaluation method in navigation. The following questions are addressed in this study: 1) What’s the relationship between the spatial familiarity with the landmark selection in navigation for individuals? 2) What is the quantitative evaluation model to connect between the spatial familiarity with the landmark salience by using the eye-tracking data?</p><p>In our experiments, we will use eye tracking to explore the influence of spatial familiarity on landmark salience. The experimental path is part of the campus area of Nanjing Normal University. The subjects are divided into novice students and senior students from Nanjing Normal University as unfamiliar and familiar groups respectively. The procedures are divided into two parts, namely, the subjects should walk the experimental path within the fixation time and then they are asked to wear VR eye tracking module a Glass DK II special for HTC view (China; https://www.7invensun.com/) to do three types of tasks about landmark picture with the characteristics of visual, semantic and structural. With this research, we expect to establish a landmark salience evaluation method through eye movement data and analyse the landmark selection rules of groups with different spatial familiarity. It is helpful to design the pedestrian navigation system based on landmarks, improve the pertinence and reliability of navigation services and reduce the cognitive load for users.</p>

An instance-based scoring system for indoor landmark salience evaluation

Navigation in a complex indoor environment can be difficult, and pedestrians may find themselves disoriented. As the featured objects of an environment, indoor landmarks play an important role in navigation. A review of the existing literature in outdoor landmark evaluation methods, however, shows that they cannot be fully applicable in any indoor environment. In this paper, an instance-based scoring system is proposed for analyzing the indicators that influence the salience of spatial objects from visual, semantic and structural aspects. An Analytic Hierarchy Process was applied to calculate landmark weights using these indicators. Two types of indoor scenes were employed as instances to verify the validity of this method, the Dongchenghui shopping mall (Nanjing, China) using a subjective questionnaire and interview method, and the headquarters of Masaryk University (Brno, Czechia) using an objective eyetracking method. Be result of the two instances showed that the proposed method was feasible.

THE INFLUENCE OF SPATIAL FAMILIARITY ON THE LANDMARK SALIENCE SENSIBILITY IN PEDESTRIAN NAVIGATION ENVIRONMENT

To contribute to a more effective design of landmark navigation guidance, this paper is concerned with the relationship between the spatial familiarity and landmark salience, which includes visual, semantic and structural attributes. The link of those two is the subjective judgment of users, which is called landmark salience sensibility. In order to explore the influence of spatial familiarity on the landmark salience sensibility, we selected two types of experimental area including campus and commercial district and four groups of experimental subject with different spatial familiarity degree. After the whole walking process, subjects are asked to draw a navigation sketch for themselves. Depending on the landmarks remaining in the sketch, we calculated the three attributes of the mean landmark salience to represent the landmark salience sensibility of each group for both paths. The result shows that with the increase of spatial familiarity, the landmark salience sensitivity is gradually reduced and the ascending order of attention degree to the attributes of the landmark salience is visual, semantic and structural salience. This conclusion is supportive to the study of landmark extraction and pedestrian guidance. Because the outdoor landmarks are analysed, we propose that in the future indoor landmarks are needed to be concerned.