Mood-Driven Colorization of Virtual Indoor Scenes.

Area of Science:

• Computer Graphics
• Human-Computer Interaction
• Virtual Reality

Background:

• Designing virtual environments to evoke specific moods is challenging due to the subjective nature of emotions.
• Current virtual reality (VR) scene design lacks automated methods for mood-specific color adjustments.

Purpose of the Study:

• To develop a novel approach for automatically adjusting the colors of virtual indoor scene textures to match a target mood.
• To enhance user immersion and emotional engagement in virtual reality experiences.

Main Methods:

• Utilized a dataset of 25,000 interior images to train a deep learning classifier for feature extraction.
• Developed an optimization process to automatically colorize virtual scenes based on extracted features and target moods.
• Conducted user studies with VR headsets to evaluate the system's efficacy.

Main Results:

• The proposed approach successfully demonstrated the ability to adjust virtual scene colors to align with target moods.
• Statistical analysis of user study data confirmed the system's effectiveness in influencing the perceived mood of virtual indoor scenes.
• The method showed promise in creating more emotionally resonant virtual reality environments.

Conclusions:

• Automated color adjustment of virtual scene textures is a viable method for mood induction in virtual reality.
• The developed deep learning-based approach offers a scalable solution for mood-aware virtual environment design.
• Further research can explore a wider range of moods and scene types.