TensorGRAF: tensorial generative radiance field

Pin-Chieh Yu; Der-Lor Way; Zen-Chung Shih

doi:10.1117/12.3016853

2 May 2024 TensorGRAF: tensorial generative radiance field

Pin-Chieh Yu, Der-Lor Way, Zen-Chung Shih

Proceedings Volume 13164, International Workshop on Advanced Imaging Technology (IWAIT) 2024; 1316403 (2024) https://doi.org/10.1117/12.3016853
Event: International Workshop on Advanced Imaging Technology (IWAIT) 2024, 2024, Langkawi, Malaysia

Abstract

3D-aware generative methods based on neural radiance fields are gaining attention. Nevertheless, they suffer from slow training and execution speeds due to volume rendering and deep neural networks. We propose using a voxel grid as the explicit representation of the radiance field, combining a shallow network to interpret the spatial features. We employ tensor decomposition to convert the voxel into axis-aligned feature vectors, reducing synthesis space complexity from O(n³) to O(n). Additionally, we leverage the well-established 2D generative adversarial network structure in our 1D feature vector generator.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Pin-Chieh Yu, Der-Lor Way, and Zen-Chung Shih "TensorGRAF: tensorial generative radiance field", Proc. SPIE 13164, International Workshop on Advanced Imaging Technology (IWAIT) 2024, 1316403 (2 May 2024); https://doi.org/10.1117/12.3016853

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available