Skip to search boxSkip to navigationSkip to main content

NeRF-To-Real Tester: Neural Radiance Fields as Test Image Generators for Vision of Autonomous Systems

Dataset:
Dataset Types
Software

Dataset Information

Date made available

2024-11-30

Publisher

ZENODO

Description

NeRF-To-Real Tester is a testing software package designed to generate test images generated from Neural Radiance Fields (NeRF), created in nerfstudio, and use NeRF-generated images to evaluate image processing Systems Under Test (SUTs). SUTs include various feature extractors like interest point detectors as well as image classifiers. This system assesses SUTs based on several metrics such as repeatability, interest point spread, L2 norm, and cosine similarity.

Cite this dataset

Brief description
Laura Weihl, Bilal Wehbe, Andrzej Wąsowski. (2024). NeRF-To-Real Tester: Neural Radiance Fields as Test Image Generators for Vision of Autonomous Systems. 10.5281/zenodo.10966288