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

Point-based methods (PBMs) can flexibly construct wavefronts of multilayer scenes and are widely used to generate computer-generated holograms (CGHs) of complex 3D scenes. However, in the holographic reconstruction of multilayer point-cloud scenes, PBMs often give rise to artifacts such as ghosting, dark bands, and ringing, which severely degrade image quality and disrupt depth continuity. To address this problem, we establish a complex-wavefront artifact model within a unified PBM-based physical framework, and systematically elucidate the underlying causes of the three types of artifacts as well as their mutual coupling relationships. On this basis, we propose an artifact-suppression-based point-cloud hologram generation method, termed Artifact Suppression Based Point Cloud Holography (ASP-HOLO). The proposed framework comprises two stages, namely target amplitude synthesis (TAS) and phase-only iterative optimization (PIO). In the TAS stage, we synthesize a set of multiplane target amplitudes with controlled artifacts and depth-of-field effects by combining multiplane angular-spectrum diffraction reconstruction with pixel-wise stitching over regions of interest (ROIs). In the PIO stage, the multiplane target amplitudes serve as optimization constraints, and a single phase-only hologram is iteratively solved via multiplane stochastic gradient descent, which effectively mitigates the overfitting problem inherent to traditional single-plane constraints. In addition, to overcome the insensitivity of conventional global metrics such as peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) to edge-localized artifacts, we further introduce an Edge-localized Artifact Metric (ELAM). ELAM quantifies both the magnitude and spatial orientation of artifact bands along the normals of geometric edges, thereby aligning more closely with the human visual perception of edge artifacts. Numerical simulations and optical experiments demonstrate that, for complex multilayer point-cloud scenes, ASP-HOLO can simultaneously suppress ghosting, dark bands, and ringing artifacts, while producing more realistic in-focus and out-of-focus visual appearances.