A Comprehensive and Detailed Study on Physically Based Rendering: From Theoretical Foundations to Real-Time and Neural Techniques

Physically Based Rendering (PBR) is a fundamental approach in computer graphics that aims to generate highly realistic images by simulating the physical behavior of light. This review paper presents an in-depth analysis of five significant research works that have shaped the evolution of PBR. The study begins with the theoretical framework that defines light transport using radiometric principles and the rendering equation [1]. It further explores Monte Carlo integration techniques, which provide numerical solutions for complex light interactions [2]. Advanced sampling methods such as Metropolis Light Transport are examined for their ability to efficiently handle difficult lighting scenarios like caustics and indirect illumination.[7] The paper also discusses real-time advancements such as Spatiotemporal Reservoir Resampling (ReSTIR), which enables high-quality rendering at interactive frame rates. [8] Additionally, neural rendering approaches are analyzed for their integration of machine learning with physically based models to improve efficiency and reduce variance. [5].  A comparative study of these methods is presented based on objectives, methodologies, and outcomes. The review highlights the transition from traditional physics-based approaches to hybrid AI-driven techniques. Finally, the paper identifies future research directions in real-time rendering, neural sampling, and GPU optimization.