Design and Evaluation of Data Annotation Workflows for CAVE-like Virtual Environments.

Area of Science:

• Computer Science
• Human-Computer Interaction
• Virtual Reality

Background:

• Data annotation is crucial for Virtual Reality (VR) applications, particularly in data analysis tasks like architectural reviews.
• Existing VR annotation systems often lack detailed explanations and evaluations of metadata handling for complex annotations.
• Interactive data annotation workflows require robust methods for creation, management, and modification of annotations.

Purpose of the Study:

• To design and evaluate an improved data annotation workflow for immersive virtual environments.
• To address the detailed handling of metadata in interactive annotation processes.
• To create an extensible workflow supporting diverse annotation types and interaction techniques.

Main Methods:

• Proposed a novel data annotation workflow design focusing on core operations: creation, review, and modification.
• Developed an extensible system for various annotation and metadata types and interaction techniques.
• Conducted a user study in a CAVE-like virtual environment comparing the proposed design against two alternatives.

Main Results:

• The proposed workflow design demonstrated good task performance in realistic annotation creation scenarios.
• User experience was positively impacted by the new annotation workflow.
• The design proved effective in facilitating complex annotation management within VR.

Conclusions:

• The developed data annotation workflow offers a significant improvement for VR applications.
• The workflow's extensibility makes it suitable for a wide range of application scenarios.
• The study validates the effectiveness of the proposed design in enhancing user performance and satisfaction.