Hybrid Mesh-Gaussian Representation for Efficient Indoor Scene Reconstruction

Huang, Binxiao; Li, Zhihao; Liu, Shiyong; Tang, Xiao; Tang, Jiajun; Lin, Jiaqi; Cheng, Yuxin; Chen, Zhenyu; Wu, Xiaofei; Wong, Ngai

doi:10.24963/ijcai.2025/128

Hybrid Mesh-Gaussian Representation for Efficient Indoor Scene Reconstruction

Binxiao Huang, Zhihao Li, Shiyong Liu, Xiao Tang, Jiajun Tang, Jiaqi Lin, Yuxin Cheng, Zhenyu Chen, Xiaofei Wu, Ngai Wong

Download BibTeX Cite

Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence. Main Track. Pages 1143-1151.

https://doi.org/10.24963/ijcai.2025/128

Keywords

Computer Vision: CV: Representation learning
Computer Vision: CV: 3D computer vision

Abstract

3D Gaussian splatting (3DGS) has demonstrated exceptional performance in image-based 3D reconstruction and real-time rendering. However, regions with complex textures require numerous Gaussians to capture significant color variations accurately, leading to inefficiencies in rendering speed. To address this challenge, we introduce a hybrid representation for indoor scenes that combines 3DGS with textured meshes. Our approach uses textured meshes to handle texture-rich flat areas, while retaining Gaussians to model intricate geometries. The proposed method begins by pruning and refining the extracted mesh to eliminate geometrically complex regions. We then employ a joint optimization for 3DGS and mesh, incorporating a warm-up strategy and transmittance-aware supervision to balance their contributions seamlessly.Extensive experiments demonstrate that the hybrid representation maintains comparable rendering quality and achieves superior frames per second FPS with fewer Gaussian primitives.

Binxiao Huang, Zhihao Li, Shiyong Liu, Xiao Tang, Jiajun Tang, Jiaqi Lin, Yuxin Cheng, Zhenyu Chen, Xiaofei Wu, and Ngai Wong. 2025. Hybrid Mesh-Gaussian Representation for Efficient Indoor Scene Reconstruction. In Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25. IJCAI, 1143–1151. https://doi.org/10.24963/ijcai.2025/128

@inproceedings{ijcai2025p128,
  title     = {Hybrid Mesh-Gaussian Representation for Efficient Indoor Scene Reconstruction},
  author    = {Huang, Binxiao and Li, Zhihao and Liu, Shiyong and Tang, Xiao and Tang, Jiajun and Lin, Jiaqi and Cheng, Yuxin and Chen, Zhenyu and Wu, Xiaofei and Wong, Ngai},
  booktitle = {Proceedings of the Thirty-Fourth International Joint Conference on
               Artificial Intelligence, {IJCAI-25}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {James Kwok},
  pages     = {1143--1151},
  year      = {2025},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2025/128},
  url       = {https://doi.org/10.24963/ijcai.2025/128},
}

TY  - CPAPER
TI  - Hybrid Mesh-Gaussian Representation for Efficient Indoor Scene Reconstruction
AU  - Huang, Binxiao
AU  - Li, Zhihao
AU  - Liu, Shiyong
AU  - Tang, Xiao
AU  - Tang, Jiajun
AU  - Lin, Jiaqi
AU  - Cheng, Yuxin
AU  - Chen, Zhenyu
AU  - Wu, Xiaofei
AU  - Wong, Ngai
T2  - Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence, IJCAI-25
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2025
SP  - 1143
EP  - 1151
DO  - 10.24963/ijcai.2025/128
UR  - https://doi.org/10.24963/ijcai.2025/128
ER  -