EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

OptiNeRF: A Spatially Optimized Neural Rendering Framework for Complex Scene Reconstruction

Xinyuan Gu, Yanbo Chang, Junyue Xia, Yue Yu, Zhen Tian, Junming Chen

Mathematics · 01.03.2026 · 14 (5) · 842

DOI: 10.3390/math14050842 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Mathematics, online erschienen am 01.03.2026. Verfasst von Xinyuan Gu und 5 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Research Fund of Macao University of Science and Technology, Science and Technology Development Fund of Macau.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Neural rendering techniques aim to generate photorealistic images and accurate 3D geometries from multi-view images but often struggle with efficiency and geometric consistency in complex or dynamic scenes. Optimized Neural Radiance Fields (OptiNeRF) addresses these challenges through several innovations. It uses spatially optimized sampling to focus on points near object surfaces, reducing computation while improving precision. Leveraging the pre-trained Marigold model, it generates depth and normal maps as geometric priors. Sampled points are processed through a hybrid network combining an MLP and a multi-resolution feature grid (MRF), capturing fine details and large-scale structures. To handle varying illumination and complex materials, OptiNeRF introduces adaptive volume rendering (AVR), dynamically adjusting light transparency and scattering. A progressive sampling strategy further focuses computation on regions with high geometric complexity. The loss function incorporates RGB, normal, depth, boundary, and lighting optimization losses, with adaptive weight modulation for geometric priors, ensuring both visual fidelity and geometric consistency even with inaccurate depth/normal estimates. Experiments on dynamic scenes show strong performance, with a PSNR of 32.10 dB, SSIM of 0.936, Chamfer distance of 1.28 × 10−3, training time of 12 h, and rendering speed of 25 FPS, demonstrating high geometric accuracy, realistic rendering, and computational efficiency over conventional methods.

Förderung

  • Research Fund of Macao University of Science and Technology
  • Science and Technology Development Fund of Macau

Zitieren

APA

Gu, X., Chang, Y., Xia, J., Yu, Y., Tian, Z., & Chen, J. (2026). OptiNeRF: A Spatially Optimized Neural Rendering Framework for Complex Scene Reconstruction. Mathematics, 14(5), 842. https://doi.org/10.3390/math14050842

Vancouver

Gu X, Chang Y, Xia J, Yu Y, Tian Z, Chen J. OptiNeRF: A Spatially Optimized Neural Rendering Framework for Complex Scene Reconstruction. Mathematics. 2026;14(5):842. doi:10.3390/math14050842

BibTeX
@article{Gu2026,
  author = {Gu, Xinyuan and Chang, Yanbo and Xia, Junyue and Yu, Yue and Tian, Zhen and Chen, Junming},
  title = {OptiNeRF: A Spatially Optimized Neural Rendering Framework for Complex Scene Reconstruction},
  journal = {Mathematics},
  year = {2026},
  volume = {14},
  number = {5},
  pages = {842},
  doi = {10.3390/math14050842}
}
Kostenlos starten →