3D Visualization of Extinct Village

3D visualization of extinct village usability was tested with the use of eye-tracking.  The experiment contained two parts – the overall views of the village and tasks, where users were searching for a particular building. Recorded eye-tracking data were statistically and graphically analyzed. The results indicate that a 3D model was the most usable type of visualization for respondents.

Methods

Procedure

Before each stimuli, the respondents had 30 seconds to read and remember the task. After that, the fixation cross was presented for 600 ms to ensure that all respondents started from the center of the stimuli. Then, the stimuli with the map was projected for 60 seconds.  The experiment contained two parts. In the first one, respondents had to look for 4 seconds at the overall views of the village without any task. For these stimuli, fixation length and count were analyzed. During the second part of the experiment, respondents had to find out a particular building in the village as fast as possible. For targets, significant buildings in the village were used (church, school, etc.). Data recorded in this part of the experiment were analyzed with the use of Areas of Interest (AOI). 

Independent variables

Three types of extinct village visualization were analysed – Cadastral Map, Orthophoto and Ortophoto with 3D model. 

Stimuli

As stimuli in the experiment, static images of the 3D model of the extinct village of Čistá were used.

 Participants

A total of 28 respondents participated in the experiment. 

Apparatus

For the study, remote eye-tracking device SMI RED 250, developed by SensoMotoric Instruments, was used. This device was operating at frequency of 120 Hz. 

 Analyses

At the beginning, overall views of the village were analyzed. Three stimuli, showing different visualizations of the South-East view of the village, were analyzed. Each image was shown to the respondents for 4 seconds. The image order was random. Two most common eye-tracking metrics were evaluated – Fixation Count and Fixation Duration. The smallest number of fixations and highest fixation length were observed for the 3D model version of the stimulus are visible. For the 3D model, a smaller number of longer fixations were observed, which indicates deeper perception. Data were statistically analyzed with the use of ANOVA and TukeyHSD. Statistically significant differences were observed for the Fixation Duration metric between Cadastral map and 3D model. These results showed that respondents’ eyes roam from place to place in the case of the cadastral map and that the 3D model was the most useful tool for the visual search.

For visual data analyses, a method called FlowMap introduced by Andrienko et al. was used. For the creation of this eye-tracking data visualization method, the software V-Analytics (a.k.a CommonGIS) developed at the Fraunhofer Institute in Germany was used. Flow Map represents results of discrete spatial and spatio-temporal aggregation of trajectories. Arrows represent multiple movement of gaze from one location to another. The thickness of arrows is derived from variable number of moves between defined voronoi polygons. Only arrows representing more than three moves are displayed. From Figure, it is visible that in the stimuli with the Cadastral map and the 3D model, the respondents’ gaze corresponds with the streets of Čistá village.

The second type of tasks in the experiment dealt with the finding of a particular building (i.e. School) in the village.  Data were analyzed with the use of Areas of Interest (AOI). These areas were marked around all important places in the stimuli, as well as around blocks of building around the square. The result of the analysis is shown in Figure.

Conclusion

 Two most often used eye-tracking metrics (Fixation Count and Fixation Duration) were investigated. A set of three stimuli (Cadastral map, Orthophoto and 3D model) with the view of the image from the South-East were investigated. The highest number of fixations was observed for the cadastral map, the lowest for the 3D model. Between these two types of visualization, the statistically significant differences were observed with the use of ANOVA. These results showed that the respondents’ eyes roamed from place to place in the case of the cadastral map and that the 3D model was the most useful tool for the visual search. Eye-tracking data were visually analyzed with the use of FlowMap method, which displays data as spatio-temporal aggregation of trajectories from all respondents. In the last part of the eye-tracking data analyses, the pair of stimuli was analyzed. In that case, the respondents had to find “school” as fast as they can. From AOI analyses it is obvious that the 3D model was more useful for this task than the orthophoto.

Outputs

Popelka, S., & Dedkova, P. (2014) Extinct Village 3D Visualization and its Evaluation with Eye-Movement Recording. Vol. 8579 LNCS. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp. 786-795).

 Downloads