Sheng Miao (苗升)

I am currently pursuing my Ph.D. at Zhejiang University, under the guidance of Prof. Yiyi Liao. Prior to this, I completed my Master's degree in Control Science and Engineering at Northwestern Polytechnical University (NWPU), advised by Prof. Xiaoxiong Liu. I received the Bachelor's degree from NWPU in 2019.

Currently, I have focused on projects involving feed-forward techniques for 3D/4D scene reconstruction and 3D generation. I serve as the conference reviewer of CVPR and NIPS.

I am looking for a full-time job in the field of 3D/4D scene reconstruction and 3D generation. Feel free to contact me at [email protected].

Email / Github

News

We release the EVolSplat4D, a feed-forward model for dynamic scene reconstruction. 😊
Nov 06, 2025: One paper is accepted to 3DV 2026.
Jun 26, 2025: One paper is accepted to ICCV 2025.
Feb 27, 2025: One paper is accepted to CVPR 2025.
Jul 1, 2024: One paper is accepted to ECCV 2024.

Selected Publications

* means equal contribution.

	EVolSplat4D: Efficient Volume-based Gaussian Splatting for 4D Urban Scene Synthesis Sheng Miao, Sijin Li, Dongfeng Bai, Pan Wang, Bingbing Liu, Yue Wang, Andreas Geiger, Yiyi Liao, Arxiv, 2026 project page / arXiv A unified feed-forward 3D Gaussian Splatting framework for static & dynamic urban scenes that achieves real-time rendering speeds. Completes scene reconstruction in ~1.3 seconds with photo-realistic quality.
	FreeFix: Boosting 3D Gaussian Splatting via Fine-Tuning-Free Diffusion Models Hongyu Zhou, Zisen Shao, Sheng Miao, Pan Wang, Dongfeng Bai, Bingbing Liu, Yiyi Liao, 3DV, 2026 arXiv A method designed to improve the rendering results of extrapolated views in 3D Gaussian Splatting, without requiring fine-tuning of diffusion models..
	Towards Depth Foundation Model: Recent Trends in Vision-Based Depth Estimation Zhen Xu^∗, Hongyu Zhou^∗, Sida Peng, Haotong Lin, Haoyu Guo, Jiahao Shao, Peishan Yang, Qinglin Yang, Sheng Miao, Xingyi He, Yifan Wang, Yue Wang, Ruizhen Hu, Yiyi Liao, Xiaowei Zhou, Hujun Bao, Arxiv, 2025 Preprint This paper surveys the evolution of deep learning architectures and paradigms for depth estimation across the monocular, stereo, multi-view, and monocular video settings. We explore the potential of these models to address existing challenges and provide a comprehensive overview of large-scale datasets that can facilitate their development.
	Vivid4D: Improving 4D Reconstruction from Monocular Video by Video Inpainting Jiaxin Huang, Sheng Miao, Bangbang Yang, Yuewen Ma, Yiyi Liao, ICCV, 2025 project page / arXiv A generative framework that addresses the challenge of reconstructing dynamic scenes from casual monocular videos with video diffusion model.
	EVolSplat: Efficient Volume-based Gaussian Splatting for Urban View Synthesis Sheng Miao, Jiaxin Huang, Dongfeng Bai, Xu Yan, Hongyu Zhou, Yue Wang, Bingbing Liu, Andreas Geiger, Yiyi Liao, CVPR, 2025 project page / arXiv A generalizable framework that predicts 3D Gaussians across multiple frames within a unified volume using convolutional networks.
	Efficient Depth-Guided Urban View Synthesis Sheng Miao^∗, Jiaxin Huang^∗, Dongfeng Bai, Weichao Qiu, Bingbing Liu, Andreas Geiger, Yiyi Liao, ECCV, 2024 project page / arXiv A generalizable NeRF model that explicitly leverages geometric priors for feed-forward inference on unbounded sparse urban scenes.
	A Visual SLAM Robust against Dynamic Objects Based on Hybrid Semantic-Geometry Information Sheng Miao, Xiaoxiong Liu, Dazheng Wei, Changze Li, ISPRS International Journal of Geo-Information, 2021, SCI Journal A method for static/dynamic image segmentation that leverages semantic and geometric modules, including optical flow residual clustering, epipolar constraint checks, semantic segmentation to tackle dynamics.