MAP3D: An explorative approach for automatic mapping of real-world eye-tracking data on a virtual 3D model

Area of Science:

• Human-Computer Interaction
• Cognitive Psychology
• Computer Vision

Background:

• Mobile eye tracking offers high ecological validity for real-world studies but presents analysis challenges.
• Current methods often reduce 3D stimuli to 2D images, losing depth information and complicating fixation mapping.
• Mapping fixations across multiple 2D reference views introduces difficulties, especially in transition zones.

Purpose of the Study:

• To introduce MAP3D, a novel approach for generating 3D models of stimuli and automatically mapping eye-tracking fixations.
• To overcome the limitations of 2D stimulus reduction and improve the accuracy of fixation data analysis in mobile eye-tracking research.
• To provide a cost-efficient, open-source solution for researchers studying visual attention in complex, real-world environments.

Main Methods:

• Development of MAP3D software to create a 3D model of the stimulus.
• Automatic mapping of recorded eye-tracking fixations onto the generated 3D model.
• Outputting fixation data as 3D coordinates (point cloud) and .csv files for further analysis.

Main Results:

• MAP3D successfully generates 3D models and automatically maps fixation data.
• The method avoids information loss associated with 2D stimulus reduction.
• Exploratory tests show promising results for accurate and efficient fixation mapping in 3D space.

Conclusions:

• MAP3D offers an innovative solution for post-hoc analysis of fixation data on 3D stimuli.
• The approach enhances the three-dimensionality of eye-tracking data analysis in mobile settings.
• MAP3D provides cost-effective, open-source tools, opening new research possibilities in visual attention studies.