Artifact-suppression-based point-cloud holography for multi-depth scenes

Area of Science:

• Computational optics and holography.
• Computer-generated holography (CGH) for 3D scene reconstruction.

Background:

• Point-based methods (PBMs) are used for CGH generation of multilayer 3D scenes.
• PBMs often produce artifacts like ghosting, dark bands, and ringing in holographic reconstructions.
• These artifacts degrade image quality and depth continuity in multilayer point-cloud scenes.

Purpose of the Study:

• To address artifacts in PBM-based CGH reconstruction of multilayer point-cloud scenes.
• To propose a novel method, ASP-HOLO, for suppressing ghosting, dark bands, and ringing artifacts.
• To enhance the visual realism and depth continuity of holographic reconstructions.

Main Methods:

• Developed a complex-wavefront artifact model within a PBM framework.
• Proposed a two-stage method: Target Amplitude Synthesis (TAS) and Phase-Only Iterative Optimization (PIO).
• TAS synthesizes multiplane target amplitudes using diffraction reconstruction and stitching; PIO uses multiplane constraints for iterative hologram solving.
• Introduced an Edge-localized Artifact Metric (ELAM) to evaluate edge artifacts, complementing PSNR and SSIM.

Main Results:

• ASP-HOLO effectively suppresses ghosting, dark bands, and ringing artifacts in complex multilayer point-cloud scenes.
• The method produces more realistic in-focus and out-of-focus visual appearances.
• Numerical simulations and optical experiments validate the artifact suppression capabilities.

Conclusions:

• ASP-HOLO provides a unified framework for artifact modeling and suppression in PBM-based CGH.
• The proposed method improves holographic reconstruction quality and visual fidelity for 3D point-cloud scenes.
• ELAM offers a more perceptually relevant metric for evaluating edge artifacts in holography.