About
Zicheng Liu is a partner research manager at Microsoft Azure AI managing the computer vision science group.
Zicheng Liu received his Ph.D. in computer science from Princeton University in 1996. He got his B.S. degree in mathematics from HuaZhong Normal University, Wuhan, China, in 1984, and his M.S. in Operations Research from the Institute of Applied Mathematics, Chinese Academy of Sciences, in 1989. Before joining Microsoft Research, he worked at Silicon Graphics, Inc. as a member of technical staff for two years, where he developed the trimmed NURBS tessellator shipped in both OpenGL and the OpenGL Optimizer.
Current research interests include vision-language learning, 3D human body and hand reconstruction, dynamic convolution, human activity recognition. He has worked on a variety of topics including Steiner trees, average case complexity, linked figure animation, and trimmed NURBS tessellation for large CAD model visualization.
Liu has served in the technical committee for many international conferences. He was a member of the Audio and Electroacoustics Committee of IEEE Signal Processing Society. He is the chair of the Multimedia Systems and Applications Technical Committee of IEEE CAS society. He is a steering committee member of IEEE Transactions on Multimedia. He is the Editor-in-Chief of Journal of Visual Communications and Image Representation, and an associate editor of Machine Vision and Applications. He served as a guest editor of IEEE Transactions on Multimedia, and a guest editor of IEEE Multimedia Magazine. He is an affiliate professor in the department of Electrical Engineering, University of Washington. He was an IEEE distinguished lecturer from 2015-2016. He is a fellow of IEEE.
Recent Work:
- DisCo: Disentangled Control for Referring Human Dance Generation in Real World. Webpage (opens in new tab) paper (opens in new tab)
- Equivariant Similarity for Vision-Language Foundation Models. ICCV 2023. paper (opens in new tab)
- Binary Latent Diffusion. CVPR2023. paper (opens in new tab)
- Reco: Region-Controlled Text-to-Image Generation. CVPR2023. paper (opens in new tab)
- NUWA-XL: Diffusion over Diffusion for eXtremely Long Video Generation. ACL 2023, Paper (opens in new tab)
- MM-ReAct: Prompting ChatGPT for Multimodal Reasoning and Action. LinkedIn (opens in new tab) , Webpage (opens in new tab), Paper, (opens in new tab)
- NUWA-Infinity: Autoregressive over Autoregressive Generation for Infinite Visual Synthesis Webpage (opens in new tab) and Paper (opens in new tab) and LinkedIn Page (opens in new tab) with more links
- CVPR 2022 Tutorial on “Recent Advances in Vision-and-Language Pre-training” (opens in new tab), 9am-5pm, June 19th, New Orleans.
- VCIP 2022 (opens in new tab)
- Azure AI Vision Image Analysis V4.0 Captioning and Dense Captioning (Public Preview) (opens in new tab)
- GIT: A Generative Image-to-Text Transformer for Vision and Language. Paper (opens in new tab)
- My team’s webpage on vision-language learning: https://www.microsoft.com/en-us/research/project/project-florence-vl/ (opens in new tab)
- OVIS: Open-Vocabulary Visual Instance Search via Visual-Semantic Aligned Representation Learning, AAAI 2022 (opens in new tab)
- An Empirical Study of GPT-3 for Few-Shot Knowledge-Based VQA, AAAI 2022 (opens in new tab)
- Playing Lottery Tickets with Vision and Language, AAAI 2022 (opens in new tab)
- VALUE: A Multi-Task Benchmark for Video-and-Language Understanding Evaluation (NeurIPS 2021) (opens in new tab)
- Azure AI Vision Spatial Analysis (opens in new tab)
- MMP-Tracking dataset: Multi-camera Multiple People Tracking Dataset, Challenge, and Workshop in conjunction with ICCV2021 (opens in new tab)
- Mesh transformer (METRO) achieves the 1st place on FreiHand leaderboard (CodaLab – Competition (opens in new tab)). Below is the paper:
- His team has reached human parity on image caption nocaps benchmark dataset. Check out paper and video demo (opens in new tab)
- Check out Azure Kinect Body Tracking SDK (opens in new tab): skeletal tracking on Azure Kinect (opens in new tab)
- MSR Action recognition datasets: https://sites.google.com/view/wanqingli/home-news (opens in new tab)
- Publications: https://zicliu.wixsite.com/mysite/publications (opens in new tab)
- His personal webpage: https://sites.google.com/view/zichengliu/home (opens in new tab)
- Google Scholar: https://scholar.google.com/citations?hl=en&user=bkALdvsAAAAJ (opens in new tab)