A Comparison of Tiny-nerf versus Spatial Representations for 3d Reconstruction

TL;DR

This paper compares tiny-NeRF neural rendering with traditional spatial representations like voxel maps, point clouds, and meshes for 3D reconstruction in robotics, focusing on memory and processing efficiency.

Contribution

It provides a systematic comparison of neural versus classical spatial representations in robotics, highlighting their respective advantages and disadvantages.

Findings

tiny-NeRF uses three times less memory than traditional methods

tiny-NeRF takes six times longer to compute the model

Neural representations offer a different trade-off between memory and processing time

Abstract

Neural rendering has emerged as a powerful paradigm for synthesizing images, offering many benefits over classical rendering by using neural networks to reconstruct surfaces, represent shapes, and synthesize novel views, either for objects or scenes. In this neural rendering, the environment is encoded into a neural network. We believe that these new representations can be used to codify the scene for a mobile robot. Therefore, in this work, we perform a comparison between a trending neural rendering, called tiny-NeRF, and other volume representations that are commonly used as maps in robotics, such as voxel maps, point clouds, and triangular meshes. The target is to know the advantages and disadvantages of neural representations in the robotics context. The comparison is made in terms of spatial complexity and processing time to obtain a model. Experiments show that tiny-NeRF requires three times less memory space compared to other representations. In terms of processing time, tiny-NeRF takes about six times more to compute the model.