NeuFG: Neural Fuzzy Geometric Representation for 3D Reconstruction

Hong, Qingqi; Yang, Chuanfeng; Chen, Jiahui; Li, Zihan; Wu, Qingqiang; Li, Qingde; Tian, Jie

Abstract

3D reconstruction from multi-view images is considered as a longstanding problem in computer vision and graphics. In order to achieve high-fidelity geometry and appearance of 3D scenes, this paper proposes a novel geometric object learning method for multi-view reconstruction with fuzzy set theory . We establish a new neural 3D reconstruction theoretical frame called neural fuzzy geometric representation (NeuFG), which is a special type of implicit representation of geometric scene that only takes value in [0, 1]. NeuFG is essentially a volume image, and thus can be visualized directly with the conventional volume rendering technique. Extensive experiments on two public datasets, i.e., DTU and BlendedMVS, show that our method has the ability of accurately reconstructing complex shapes with vivid geometric details, without the requirement of mask supervision. Both qualitative and quantitative comparisons demonstrate that the proposed method has superior performance over the state-of-the-art neural scene representation methods. The code will be released on GitHub soon.

Citation

Hong, Q., Yang, C., Chen, J., Li, Z., Wu, Q., Li, Q., & Tian, J. (2024). NeuFG: Neural Fuzzy Geometric Representation for 3D Reconstruction. IEEE Transactions on Fuzzy Systems, https://doi.org/10.1109/TFUZZ.2024.3447088

• Journal Article Type: Article
• Acceptance Date: Aug 18, 2024
• Online Publication Date: Aug 21, 2024
• Publication Date: Jan 1, 2024
• Deposit Date: Aug 23, 2024
• Publicly Available Date: Sep 2, 2024
• Journal: IEEE Transactions on Fuzzy Systems
• Print ISSN: 1063-6706
• Publisher: Institute of Electrical and Electronics Engineers
• Peer Reviewed: Peer Reviewed
• DOI: https://doi.org/10.1109/TFUZZ.2024.3447088
• Keywords: multi-view images, Surface reconstruction , Geometric object learning , Fuzzy sets
• Public URL: https://hull-repository.worktribe.com/output/4790803