Selecting Tangible Media for Immersive Exploration of Volumetric Scientific Data

Paper Title

Selecting Tangible Media for Immersive Exploration of Volumetric Scientific Data

Publication Info

• Topic area: Tangible media design for immersive scientific data exploration
• Keywords: Tangible interaction, augmented reality, volumetric data, 1D media, 2D media, 3D media, scientific visualization, user study, design space, interaction techniques

Background and Problem

• Problem / challenge: Designers lack clear guidance on selecting appropriate physical dimensionality (1D, 2D, or 3D) for tangible media in immersive scientific data exploration tasks. Existing research does not systematically compare how dimensionality impacts user interaction.
• Significance: Tangible media can bridge the gap between physical and digital interactions, reducing cognitive load and improving precision in scientific data exploration. Clear guidelines are necessary to optimize task performance and usability.
• Motivation and related work: Prior studies have explored specific tangible forms (e.g., cubes, globes) or interaction techniques but have not systematically evaluated the role of dimensionality in scientific visualization. This paper addresses this gap by empirically comparing 1D, 2D, and 3D tangible media.

Solution

• Proposed approach: A design space that structures the relationship between tangible media dimensionality, interaction techniques, and scientific visualization tasks, combined with a prototype system for empirical evaluation.
• Novelty:
  1. Comprehensive empirical evaluation of 1D, 2D, and 3D tangible media for volumetric data exploration.
  2. A structured design space linking media dimensionality, interaction techniques, and visualization tasks.
  3. Evidence-based guidance for selecting tangible media based on task requirements.
• Procedure and key techniques:
  1. Develop a design space with two dimensions: data visualization space (task requirements) and interaction space (gestures and manipulations).
  2. Build a prototype system implementing seven core visualization tasks (e.g., navigation, slicing, filtering).
  3. Conduct a user study with 35 participants, comparing performance and preferences across 1D, 2D, and 3D media for each task.

Results

• Concrete findings:
  • 3D media were preferred for tasks requiring volumetric manipulation (e.g., navigation, rotation).
  • 2D media excelled in planar tasks (e.g., Focus+Context, Overview+Details) and parameter adjustments.
  • 1D media were favored for single-parameter tasks (e.g., histogram filtering) due to simplicity and precision.
• Advantage over baselines: Demonstrated task-specific advantages of each medium, providing empirical evidence for selecting dimensionality rather than relying on intuition.
• Experiments / evaluation:
  • Seven tasks evaluated with 35 participants using 1D (pen), 2D (card), and 3D (box) media.
  • Metrics: task completion time (TCT), accuracy, and subjective ratings (intuitiveness, comfort, ease of operation, effectiveness).
  • Results showed task-dependent preferences and trade-offs in performance and usability.
• Limitations and future work:
  • Limited to voxel-based volumetric data; future work could explore 3D point clouds or surface meshes.
  • Tangible media tested were canonical shapes; alternative geometries and materials could be explored.
  • Adaptive, context-aware tangible interfaces could enhance flexibility and usability.

Summary

This paper investigates how the dimensionality of tangible media (1D, 2D, 3D) affects user performance and experience in immersive scientific data exploration. A design space was developed to link task requirements with media affordances, and a prototype system was used to evaluate seven core visualization tasks. The results showed that 3D media are optimal for volumetric manipulation, 2D media for planar and parameter-based tasks, and 1D media for single-parameter adjustments. These findings provide evidence-based guidance for selecting tangible media, advancing the design of immersive scientific visualization interfaces.