Computer Vision · Computational Imaging · 3D/4D Reconstruction

Cheng Peng

I work on computer vision and computational imaging, especially high-quality 3D/4D reconstruction, novel view synthesis, generative modeling, and practical image-based systems for real-world environments.

Open JHU in 3D Google Scholar LinkedIn

Research

Vision algorithms with engineering weight.

My research focuses on high-quality, high-resolution reconstruction from visible-spectrum signals such as neural radiance fields and 3D Gaussian splatting, as well as invisible-spectrum data such as CT and MRI. I also work on segmentation, detection, re-identification, and medical image analysis.

I enjoy work that combines new scientific progress with grounded engineering: methods that are mathematically coherent, computationally practical, and useful outside curated benchmarks.

3D/4D Reconstruction Novel View Synthesis Gaussian Splatting Generative Modeling Medical Imaging Computational Imaging

Recruitment

Prospective students and collaborators.

I will join the University of Virginia in Spring 2026 as an Assistant Professor at the School of Data Science. I am actively recruiting prospective PhD, Master, and Bachelor students for Fall 2026.

I mainly work in computer vision and computational imaging, with applications in site construction, robotic surgery, medical imaging, sports, and related domains.

Prospective students can review the UVA PhD program and fill out the recruitment form here.

Recent Activities

Updates synced from the original site.

  • 2026/6/18 Two ECCV papers (CAM3R, SyncFix) were accepted, congrats to Namitha and Deming!
  • 2026/3/13 Glad to receive the NVIDIA Academic Grant Award; thankful for the support from Pengfei Guo, Daguang Xu, Max Li, and Ming-Yu Liu!
  • 2026/2/21 Three CVPR papers (Scenes as Tokens, WRIVINDER, and SPIDER) were accepted, congrats to Yutao, Chandrakanth, and Zhimin!
  • 2026/1/26 Two ICLR papers (ULTRA-360 and WoW!) were accepted, congrats to Xijun and Jiahan!
  • 2025/10/19 One IEEE JSTSP (IF: 13.7) paper (FusionRF) was accepted, congrats to Michael!
  • 2025/9/18 One NIPS paper (MS-GS) was accepted, congrats to Deming!
  • 2025/6/12 Two CVPR papers (SPARS3R, Geometry Splatting) were accepted, congrats to Yutao, Yuxiang, and Kaiwen!
  • 2025/6/12 Glad to have been selected as CVPR 2025's Outstanding Reviewer!
  • 2025/4/24 One ICLR paper (Genex) was accepted, congrats to Taiming!
  • 2025/3/10 Glad to receive the NVIDIA Academic Grant Award; thankful for the support from Ming-Yu Liu and Yogesh Balaji! The proposal was written in collaboration with Jieneng Chen and Prof. Alan Yuille.
  • 2025/2/28 Organized a WACV workshop on Unconstrained Large-scale Three-dimensional Reconstruction and Rendering across Altitutdes (ULTRRA), thankful for Ashwini Deshpande, Zexiang Xu, Noah Snavely, Deva Ramanan, and Yohann Cabon as guest speakers!
  • 2025/2/28 Two WACV papers (GaitContour, VILLS) were accepetd, congrats to Yuxiang and Siyuan!
  • 2024/12/16 One NIPS paper (LP-3DGS) was accepted, congrats to Zhaoliang!
  • 2024/9/29 One ECCV paper (BAGS) was accepted, congrats to myself and Yutao!
  • 2024/9/15 One IJCB paper (GADER) was accepted with the Best Biometrics Student Paper Award, congrats to Yuxiang!
  • 2024/7/1 Joined Johns Hopkins University as an Assistant Research Professor :)

AI Venue Publications

Synced from Google Scholar, with current acceptance updates.

Students

Current advising students listed on the original site.

  • Yuxiang GuoJHU ECE PhD
  • Deming LiJHU ECE PhD
  • Yutao TangJHU BME PhD
  • Zhaoliang ZhangJHU ECE PhD
  • Siyuan HuangJHU ECE PhD
  • Zhimin ZhaoJHU ECE PhD

Collaborations

Large-scale programs and research partners.

I am supported by and collaborate across multiple research programs, particularly IARPA WRIVA and ONR MURI. I am the project manager and a co-PI in WRIVA, a multi-institute effort focused on robust, image-only 3D site modeling.

I have also contributed to IARPA BRIAR, and collaborate with researchers across UCSD, UCSB, UMD, Rice, JHU, NVIDIA, and other institutions.

Explore one current spatial reconstruction project: JHU in 3D.