WebLFR: An interactive light field renderer in web browsers

Xiaofei Ai,Yigang Wang,Ying Wu,Simin Kou

DOI: https://doi.org/10.1007/s11042-024-18526-z

IF: 2.577

2024-02-24

Multimedia Tools and Applications

Abstract:Traditional light field rendering techniques, relying on dense sampling data, can faithfully reproduce photo-realistic and view-dependent effects such as reflection and refraction. However, they consume substantial memory resources for rapid access to light rays. Light field compression can significantly reduce disk storage requirements, but it cannot mitigate the memory overhead associated with rendering. We propose a novel rendering approach that combines the strengths of light field techniques and overcomes their memory limitations. By computing light ray parameters during the rasterization stage, our method enables rapid acquisition of light ray coordinates corresponding to pixels in the image plane. By incorporating a texture compression technology to compress light field images, our method can dramatically reduce memory consumption for dense light field rendering. Experimental evaluation on multiple light field scenes demonstrates that our method can effectively minimize memory usage to 12.5% compared to the baseline technique, without compromising visual quality. Moreover, it enables real-time rendering of photo-realistic and view-dependent effects within standard WebGL-supported browsers.

computer science, information systems, theory & methods,engineering, electrical & electronic, software engineering