What Draws Your Attention First? An Attention Prediction Model Based on Spatial Features in Virtual Reality

Area of Science:

• Human-Computer Interaction
• Computer Vision
• Virtual Reality

Background:

• Understanding visual attention is crucial for effective human-computer interaction (HCI), particularly in immersive technologies like virtual reality (VR) and augmented reality (AR).
• The influence of 3D spatial attributes on visual attention in these environments remains an area requiring further quantitative exploration.
• Existing models often lack the specificity needed for complex 3D virtual environments.

Purpose of the Study:

• To quantitatively model the probability of first attention between two stimuli in a virtual reality environment.
• To develop predictive models for user visual attention based on stimulus spatial properties.
• To provide tools for VR designers to enhance user engagement and guidance in 3D content.

Main Methods:

• An experiment was conducted in VR to collect a gaze dataset of users viewing synthetic scenes with varying spatial configurations of two spheres.
• A probability model was formulated using view-specific stimulus attributes, including eccentricity and visual angle size.
• Two machine learning models were trained on the gaze dataset to predict visual attention probability distributions.

Main Results:

• The developed models demonstrated the ability to predict user preferences for visual stimuli within VR scenes.
• Model performance was evaluated across two distinct synthetic VR environments, validating their predictive capabilities.
• The study successfully created and released a valuable gaze dataset and source code for VR attention research.

Conclusions:

• The developed attention prediction models are applicable to two-foreground-object scenarios common in VR/AR content design.
• These models can assist VR designers in optimizing visual attention for applications such as storytelling and scene guidance.
• The released dataset and code facilitate further research into visual attention mechanisms in immersive environments.