Three-dimensional computer holography enabled from a single 2D image.

Area of Science:

• Optics and Photonics
• Computer Vision
• Computational Imaging

Background:

• Generating high-quality computer-generated holograms (CGH) for true 3D scenes typically requires extensive 3D data acquisition using specialized devices or complex rendering techniques.
• Existing methods for real-world CGH computation often rely on costly depth cameras and cumbersome 3D data processing.

Purpose of the Study:

• To propose a computational framework for generating CGH from a single 2D image.
• To enable straightforward 3D holographic display creation without the need for specialized 3D data acquisition.
• To outperform existing CGH computation methods in terms of simplicity and cost-effectiveness.

Main Methods:

• A deep view synthesis neural network was developed to generate light-field content from a single input image.
• A ray-wave algorithm was employed to convert the synthesized light-field data into the diffractive wavefront required for hologram generation.
• The framework utilizes a 2D-to-3D wavefront conversion approach.

Main Results:

• The proposed method successfully generates 3D CGH from single 2D images, enabling straightforward holographic display creation.
• Experimental demonstrations confirmed the ability to reconstruct both indoor and outdoor scenes using phase-only CGH generated from single images.
• The approach significantly simplifies the process compared to methods requiring depth cameras and complex 3D rendering.

Conclusions:

• A novel and efficient computational framework for generating 3D CGH from single images has been successfully developed.
• This method offers a cost-effective and accessible solution for creating holographic displays from readily available 2D image content.
• The findings pave the way for more widespread adoption of holographic technologies by reducing data acquisition and processing barriers.