Explore · 每周精选
发现最受关注的研究论文,追踪研究趋势,订阅感兴趣的期刊与关键词。
33 篇论文 · 按点赞排序
Xiangrui Liu, Xinju Wu, Pingping Zhang +3 authors
The method CompGS improves Gaussian splatting by compressing 3D scene data with a hybrid primitive structure and rate-constrained optimization, maintaining model accuracy and rendering quality.
Chieh Hubert Lin, Changil Kim, Jia-Bin Huang +5 authors
The proposed framework addresses limitations in NeRF editing by tempering diffusion model stochasticity and mitigating textural shifts, leading to state-of-the-art NeRF inpainting results.
Yuqun Wu, Jae Yong Lee, Chuhang Zou +2 authors
A patch-based approach enhances the geometric accuracy and view synthesis of NeRFs by leveraging surface normals and depth predictions, showing significant performance improvements on MVS benchmarks.
北京市昌平区探索星信息技术及软件开发工作室
京ICP备2026059466号