Complex-Valued Holographic Radiance Fields

2025-06-10

AI Summary

This work addresses the challenge of modeling physically renderable 3D representations for holographic display. We propose the first Complex-Valued 3D Gaussian Splatting framework, which jointly encodes both amplitude and phase of light waves directly in 3D space. Departing from conventional intensity-based intermediate representations, our method performs end-to-end optimization of complex-valued Gaussian primitives to achieve geometric–physical co-alignment of the holographic radiance field. By integrating RGB-D multi-view supervision with direct wavefield rendering, we eliminate the need for iterative hologram re-optimization. Compared to state-of-the-art methods, our approach achieves 30×–10,000× speedup while maintaining competitive PSNR and SSIM performance. This significantly enhances both the efficiency of holographic content generation and the physical fidelity of the reconstructed wavefield.

Technology Category

• Computer Vision: 3D Computer Vision
• Search and Optimization: Mixed Discrete/Continuous Search
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Modeling the full properties of light, including both amplitude and phase, in 3D representations is crucial for advancing physically plausible rendering, particularly in holographic displays. To support these features, we propose a novel representation that optimizes 3D scenes without relying on intensity-based intermediaries. We reformulate 3D Gaussian splatting with complex-valued Gaussian primitives, expanding support for rendering with light waves. By leveraging RGBD multi-view images, our method directly optimizes complex-valued Gaussians as a 3D holographic scene representation. This eliminates the need for computationally expensive hologram re-optimization. Compared with state-of-the-art methods, our method achieves 30x-10,000x speed improvements while maintaining on-par image quality, representing a first step towards geometrically aligned, physically plausible holographic scene representations.

Problem

• Modeling light amplitude and phase in 3D representations
• Optimizing 3D scenes without intensity-based intermediaries
• Achieving fast holographic rendering with complex-valued Gaussians

Innovation

• Complex-valued Gaussian primitives for 3D scenes
• RGBD multi-view images optimize holographic representation
• 30x-10,000x speedup without quality loss