|
Xiyi Chen I am a second-year PhD student in Computer Science at the University of Maryland, College Park, advised by Prof. Ming C. Lin. I also work closely with Prof. Ruohan Gao. Before that, I obtained my master's degree in Computer Science at ETH Zürich, where I worked on human avatar synthesis and human mesh recovery, advised by Dr. Sergey Prokudin and Prof. Siyu Tang. I obtained my bachelor's degree in Computer Science also at Maryland. Back then, I worked with Prof. David Jacobs on surveillance-quality face recognition. My research interests lie at the intersection of computer vision and computer graphics, focusing on digital human avatars and 3D/4D scene reconstruction and generation. Email / Github / Google Scholar / LinkedIn / Gallery |
|
News
|
Publications |
|
 Scale3D: Scalable 3D Reconstruction and Understanding
Phuc Nguyen, Xiyi Chen, Dongki Jung, Anshul Rai, Guan-Ming Su, Dinesh Manocha, Ming C. Lin In Submission project page We propose Scale3D, a novel framework that jointly leverages geometry and perception for scalable, globally consistent 3D reconstruction and ID-consistent scene understanding from long sequences, improving reconstruction, 3D instance segmentation, and panoptic lifting. |
|
Derong Jin*, Xiyi Chen*, Ming C. Lin, Ruohan Gao CVPR, 2026 project page / arXiv / code We propose the first framework to generate an explorable 3D scene and semantically grounded spatial ambisonic audio from a single image for free-viewpoint audio-visual rendering. |
|
Xiyi Chen, Shaofei Wang, Marko Mihajlovic, Taewon Kang, Sergey Prokudin, Ming C. Lin ArXiv, 2025 project page / arXiv We propose a unified framework for clothed human reconstruction, SMPL-X estimation, and novel view synthesis from sparse-view, uncalibrated human images. |
|
Yutong Chen, Marko Mihajlovic, Xiyi Chen, Yiming Wang, Sergey Prokudin, Siyu Tang ICLR, 2025 (Spotlight Presentation) project page / arXiv / code We analyze the performance of novel view synthesis methods in challenging out-of-distribution (OOD) camera views and introduce SplatFormer, a data-driven 3D transformer designed to refine 3D Gaussian Splatting primitives for improved quality in extreme camera scenarios. |
|
Xiyi Chen, Marko Mihajlovic, Shaofei Wang, Sergey Prokudin, Siyu Tang CVPR, 2024 project page / arXiv / code We propose the first diffusion model to enable the creation of fully 3D-consistent, animatable, and photorealistic human avatars from a single image of an unseen subject. |
Projects |
|
Semester Project at Computer Vision and Learning Group report / slides / code We improve the SMPLify-X optimization pipeline by applying keypoints blending and a stronger pose prior for robust human mesh inference on partially-observed human images. |
|
Final project for Digital Humans report / code We reproduce the imitation tasks in Deepmimic with a curriculum training strategy to extend our algorithm’s applicability to various biped robots with different shapes, masses, and dynamics models. |
Services
|
Awards
|
|
Template credits: Jon Barron |