A Perception-driven Hybrid Decomposition for Multi-layer Accommodative Displays

Area of Science:

• Computer Vision
• Display Technology
• Human-Computer Interaction

Background:

• Multi-focal plane and multi-layered light-field displays aim to replicate real-world visual cues.
• Current methods for driving these displays involve computationally expensive optimizations or inaccurate linear blending.
• Existing techniques struggle with real-time performance or fidelity, especially at occlusion boundaries and on glossy surfaces.

Purpose of the Study:

• To develop a perception-based hybrid decomposition technique for light-field displays.
• To achieve both real-time performance and high-fidelity image reconstruction.
• To reduce computational cost by selectively applying optimizations.

Main Methods:

• A perception-informed analysis and model are developed to determine optimal decomposition strategies.
• Expensive optimizations are applied selectively to perceptually critical regions (e.g., depth discontinuities).
• Less computationally intensive linear blending is used for other regions.

Main Results:

• The proposed hybrid technique successfully combines real-time performance with high-fidelity results.
• Selective application of optimizations significantly reduces computational load while maintaining visual quality.
• User experiments on a custom multi-plane display validate the effectiveness of the method.

Conclusions:

• The perception-based hybrid decomposition offers a superior approach for driving light-field displays.
• This method addresses the limitations of existing techniques, enabling practical real-time applications.
• Future work may involve further refinement of the perceptual model and hardware integration.