Usability of multiviewpoint images for spatial interaction in projection-based display systems

Area of Science:

• Human-Computer Interaction
• Computer Graphics
• Virtual Reality

Background:

• Large screen projection-based stereoscopic displays are often shared by multiple users.
• Existing interaction methods like head-tracking require special hardware and may not scale well for group use.
• Direct first-person interaction without head-tracking is often impractical in these shared environments.

Purpose of the Study:

• To compare the usability and interaction performance of multi-viewpoint images against non-head-tracked and head-tracked methods.
• To evaluate multi-viewpoint interaction for ray-casting selection and in-hand object manipulation tasks.
• To determine the effectiveness of multi-viewpoint images in shared projection-based stereoscopic display systems.

Main Methods:

• A qualitative and quantitative study was conducted.
• Compared interaction performance using multi-viewpoint images, non-head-tracked interaction, and head-tracked interaction.
• Evaluated two interaction tasks: ray-casting selection and in-hand object manipulation.

Main Results:

• Multi-viewpoint image interaction showed performance comparable or superior to head-tracked interaction.
• Ray-casting selection with multi-viewpoint images was up to 10% faster than head-tracked interaction.
• In-hand object manipulation using multi-viewpoint images was only 6% slower than head-tracked interaction.

Conclusions:

• Multi-viewpoint images provide a viable and effective interaction method for shared large screen stereoscopic displays.
• This approach eliminates the need for special hardware, making it scalable for multiple colocated users.
• Multi-viewpoint interaction offers a practical alternative to head-tracking, significantly improving usability in group scenarios.