About Me

I'm a final year Master student at UCLA, working closely with Prof. Bolei Zhou. I did research at Intelligent Vision Group (IVG), Tsinghua University, under the guidance of Prof. Jiwen Lu and Dr. Guangyi Chen. I also worked at SenseTime Research as an intern. In December 2020, I received B.S. degree in Computer Science and Mathematics at the University of Maryland. My research interest is in Computer Vision, especially Diffusion Models and Generative Modeling.

News

2022-10: I join OPPO US Research as a Research Intern.
2022-03: Two papers (1 oral and 1 poster) accepted to CVPR'22!
2022-01: I move to LA for Master degree at UCLA.
2022-03: Start an internship at SenseTime Research!
2021-03: One paper accepted by TIP'21!

Education

University of California, Los Angeles, CA, USA

Master of Engineering, Expected: August 2022
TA for computer vision course

University of Maryland at College Park, MD, USA

Bachelor of Science in Computer Science
Bachelor of Science in Mathematics

Publications

[1] Stochastic Trajectory Prediction via Motion Indeterminacy Diffusion
Tianpei Gu*, Guangyi Chen*, Junlong Li, Chunze Lin, Yongming Rao, Jie Zhou , Jiwen Lu
IEEE/CVF Conference on Computer Vision and Patter Recognition (CVPR), 2022, (Poster) dise

[Arxiv] [Code]

We propose a new framework to formulate the trajectory prediction task as a reverse process of motion interminacy diffusion, in which we progressively discard indeterminacy from all the walkable areas until reaching the desired trajectory.

[2] Bailando: 3D Dance Generation by Actor-Critic GPT with Choreographic Memory
Li Siyao, Wenjiang Yu, Tianpei Gu, Chunze Lin, Quan Wang, Chen Qian, Chen Chang Loy, Ziwei Liu
IEEE/CVF Conference on Computer Vision and Patter Recognition (CVPR), 2022, (Oral) dise

[Arxiv] [Code]

[3] Person Re-identification via Attention Pyramid
Guangyi Chen, Tianpei Gu, Jiwen Lu , Jin-An Bao, Jie Zhou
IEEE Transactions on Image Processing (TIP), 2021
dise

[PDF] [Supp] [Code]

We propose attention pyramid networks by the "split-attend-merge-stack" principle to jointly learn the attentions under different scales and obtain superior performance on many person re-identification datasets.

Industry Experience

SenseTime Research

Computer Vision Research Intern

Build an end-to-end image generation pipeline with StyleGAN to produce massive high-quality stylized image with model blending.
Main contributor of the product which transferring human face image into multiple style (now 10+).
Develop a pipeline of “Inversion-Editing-Stylization” for human face, which the editing part can receive a certain attribute selection or a text prompt.
Assisted to implement the project of music to dance generation

Service

Reviewer for CVPR2022, ECCV2022, CVPR2023

Professional Skills

Programming Language (ranked by proficiency): Python, Java, C, C++, Matlab, Git, Shell

Deep Learning Framework: PyTorch, Keras, Tensorflow