Xinqi Liu

Xinqi Liu | 刘鑫琦

Hi, this is Xinqi Liu's home page. I am currently a senior researcher at Department of Computer Vision Technology (VIS), Baidu Inc. I obtained my Ph.D. degree from School of Mechanical Engineering, Zhejiang University , advised by Prof. Jituo Li.

My research focuses on 3D AIGC, 3D Reconstruction and 3D Digital Human. My main goal is reconstruct robust and vivid human bodies, clothing and scenes from the real world in a low-cost way.

Email: liuxinqi7@126.com

Research

	GVA: Reconstructing Vivid 3D Gaussian Avatars from Monocular Videos Xinqi Liu, Chenming Wu, Jialun Liu, Xing Liu, Jinbo Wu, Chen Zhao, Haocheng Feng, Errui Ding, Jingdong Arxiv Preprint, 2024 Paper, Project A 3D Gaussian human body reconstruction method that supports both body and hand driving from monocular RGB videos.
	TexRO: Generating Delicate Textures of 3D Models by Recursive Optimization Jinbo Wu, Xing Liu, Chenming Wu, Xiaobo Gao, Jialun Liu, Xinqi Liu, Chen Zhao, Haocheng Feng, Errui Ding, Jingdong Wang Arxiv Preprint, 2024 Paper, Project A novel method for generatingdelicate textures of a known 3D mesh by optimizing its UV texture.
	TexOct: Generating Textures of 3D Models with Octree-based Diffusion Jialun Liu, Chenming Wu, Xinqi Liu, Xing Liu, Jinbo Wu, Haotian Peng, Chen Zhao, Haocheng Feng, Jingtuo Liu, Errui Ding IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024 Paper A texture generative model for 3D objects based on octree structure.
	Reconstructing Complex Shaped Clothing from a Single Image with Feature Stable Unsigned Distance Fields Xinqi Liu, Jituio Li, Guodong Lu IEEE Transactions on Visualization and Computer Graphics (TVCG), 2024 Paper, Project A clothing reconstruction method based on an unsigned distance field, which can reconstruct clothing with complex shapes and contours from a single image.
	Learning Pose Controllable Human Reconstruction with Dynamic Implicit Fields from a Single Image Jituio Li, Xinqi Liu, Guodong Lu IEEE Transactions on Visualization and Computer Graphics (TVCG), 2024 Paper, Project A novel implicit representation to generate reposed reconstruction human body from a single image and target pose.
	Modeling Realistic Clothing from a Single Image under Normal Guide Xinqi Liu, Jituio Li, Guodong Lu IEEE Transactions on Visualization and Computer Graphics (TVCG), 2023 Paper, Project A robust clothing modeling method to generate a 3D clothing model with visually consistent clothing style and wrinkles distribution from a single RGB image using normal guidance.
	Wrinkles Realistic Clothing Reconstruction using a combined Implicit and Explicit Method Xinqi Liu, Jituio Li, Guodong Lu Computer Aided Design (CAD), 2023 Paper, Project We propose a new method that combines implicit and explicit ideas to reconstruct wrinkles realistic 3D clothing model and texture results.
	Generating High-fidelity Texture in RGB-D Reconstruction using Patches Density Regularization Xinqi Liu, Jituio Li, Guodong Lu Computer Aided Design (CAD), 2023 Paper, Project A simple but effective regularization term is designed to deal with the texture patches fragmentation problem in texture mapping methods.
	Robust and Automated Body and Clothing Reconstruction from a Single RGB Image Xinqi Liu, Jituio Li, Guodong Lu Computers & Graphics, 2023 Paper, Project An automatic body and clothing reconstruction method to generate the human body with style-matched and texture-realistic clothing results from a single image.
	Reconstruction of Colored Soft Deformable Objects Based on Self-Generated Template. Jituo Li, Xinqi Liu, Haijing Deng, Tianwei Wang, Guodong Lu, Jin Wang Computer Aided Design (CAD), 2022 Paper, Project A new method to reconstruct a deformable soft object with complete geometry and consistent texture by introducing an incremental-completion self-generated template (SGT).
	Improving RGB-D based 3D Reconstruction by Combining Voxels and Points Xinqi Liu, Jituio Li, Guodong Lu, Dongliang Zhang, Shihai Xing The Visual Computer, 2022 Paper A flexible 3D reconstruction framework that adopts a new data structure combining both voxels and points to significantly improve the reconstruction accuracy.

......