New era in image augmentation – neural radiance fields

Abstract

In this article, a new method of image augmentation using Neural Radiance Fields is being considered. NeRF has significant advantages in generating photo-realistic images with high resolution and fidelity, and it can be used for a wide range of applications, including 3D rendering, virtual or augmented reality, and robotics. Despite its computational cost and limitations in some types of scenes, NeRFs are becoming more and more popular in many fields of computer vision, with ongoing research to improve its limitations. Also, this work describes a practical example of how it can preprocess and augment an existing dataset to use it for a specific aim