A High-Speed CGH Calculation Method for Mirror Images on Bézier Surfaces using Optical Path Length Minimization
Kodai Ono, Seok Kang, Yuji Sakamoto
TL;DR
This paper tackles the challenge of rendering reflections from free-form Bézier mirrors in CGH, where existing methods either cost excessive computation or fail to handle curved surfaces. It introduces a Fermat's-principle–based formulation that minimizes the optical-path length $L(s,t)$ on the Bézier surface and solves for the reflection point with Newton's method, using proximity-based initialization and GPU-accelerated ray tracing. The approach yields substantial speedups (over 100× versus mirror-subdivision and over 2,800× versus Bézier clipping) while maintaining high fidelity, and it naturally extends to multiple reflections. Experimental validations with cylindrical and parabolic mirrors, motion parallax, and multi-mirror scenarios demonstrate accurate image depth, parallax, and scalable computation, enabling potential real-time holographic rendering of complex scenes.
Abstract
Rendering reflections in curved mirrors is crucial for enhancing the realism in computer-generated hologram (CGH), yet it poses a fundamental challenge due to the unique computational principles of CGH. Conventional methods using Bézier clipping are computationally prohibitive, and a previously proposed mirror surface subdivision method suffered from the computation time increasing with mirror curvature. To address these limitations, this paper proposes a novel calculation method based on Fermat's principle that directly and efficiently determines the reflection point by minimizing the optical path length from a point light source to a hologram pixel via the mirror surface, using Newton's method for optimization. Experimental results demonstrate that this method significantly reduces computation time compared to previous approaches. Furthermore, it enables the rendering of multiple reflections from several mirrors, a capability that was challenging for conventional techniques.
