LookCloser🔍: Frequency-aware Radiance Field for Tiny-Detail Scene

CVPR 2025
Xiaoyu Zhang*1, Weihong Pan*2, Chong Bao2, Xiyu Zhang, Xiaojun Xiang, Hanqing Jiang, Hujun Bao
1SenseTime Research, 2Zhejiang University
*Indicates Equal Contribution
Paper arXiv Code Multi-frequency Dataset

Abstract

    Humans perceive and comprehend their surroundings through information spanning multiple frequencies. In immersive scenes, people naturally scan their environment to grasp its overall structure while examining fine details of objects that capture their attention. However, current NeRF frameworks primarily focus on modeling either high-frequency local views or the broad structure of scenes with low-frequency information, limited to balance both. We introduce FA-NeRF, a novel frequency-aware framework for view synthesis that simultaneously captures the overall scene structure and high-definition details within a single NeRF model.


Paper Preview



Real-time rendering and roaming, observing the overall scene structure and fine details.

Overview

Paper Preview

    We propose a 3D frequency quantification method that analyzes the scene’s frequency distribution, enabling frequency-aware rendering. Our framework incorporates a frequency grid for fast convergence and querying, a frequency-aware feature re-weighting strategy to balance features across different frequency contents.


BibTeX

@misc{zhang2025lookcloserfrequencyawareradiancefield,
    title={LookCloser: Frequency-aware Radiance Field for Tiny-Detail Scene}, 
    author={Xiaoyu Zhang and Weihong Pan and Chong Bao and Xiyu Zhang and Xiaojun Xiang and Hanqing Jiang and Hujun Bao},
    year={2025},
    eprint={2503.18513},
    archivePrefix={arXiv},
    primaryClass={cs.CV},
    url={https://arxiv.org/abs/2503.18513}
}