Evaluation of strategies for efficient rate-distortion NeRF streaming

TL;DR

This paper compares pixel-based and neural network parameter-based NeRF streaming strategies, analyzing their rate-distortion performance and highlighting the superior efficiency of the NN parameter approach for scalable 3D scene streaming.

Contribution

It provides a comprehensive evaluation of two NeRF streaming strategies, revealing the advantages of neural network parameter-based transmission in terms of efficiency and scalability.

Findings

NN parameter-based strategy outperforms pixel-based in efficiency

Trade-offs between complexity and performance are analyzed

Neural network parameter transmission is suitable for scalable streaming

Abstract

Neural Radiance Fields (NeRF) have revolutionized the field of 3D visual representation by enabling highly realistic and detailed scene reconstructions from a sparse set of images. NeRF uses a volumetric functional representation that maps 3D points to their corresponding colors and opacities, allowing for photorealistic view synthesis from arbitrary viewpoints. Despite its advancements, the efficient streaming of NeRF content remains a significant challenge due to the large amount of data involved. This paper investigates the rate-distortion performance of two NeRF streaming strategies: pixel-based and neural network (NN) parameter-based streaming. While in the former, images are coded and then transmitted throughout the network, in the latter, the respective NeRF model parameters are coded and transmitted instead. This work also highlights the trade-offs in complexity and performance, demonstrating that the NN parameter-based strategy generally offers superior efficiency, making it suitable for one-to-many streaming scenarios.