Generative Densification: Learning to Densify Gaussians for High-Fidelity Generalizable 3D Reconstruction

Seungtae Nam; Xiangyu Sun; Gyeongjin Kang; Younggeun Lee; Seungjun Oh; Eunbyung Park

doi:10.1109/cvpr52734.2025.02485

CVPR2025顶会

Generative Densification: Learning to Densify Gaussians for High-Fidelity Generalizable 3D Reconstruction

Seungtae Nam, Xiangyu Sun, Gyeongjin Kang, Younggeun Lee, Seungjun Oh, Eunbyung Park

arXiv DOI 出版方

2025年份

3顶会引用

摘要

hanced generalization. Experimental results on both objectlevel and scene-level reconstruction tasks demonstrate that our method outperforms state-of-the-art approaches with comparable or smaller model sizes, achieving notable improvements in representing fine details.