Peng Chen

Peng Chen

I am a second-year master student at Institute of Software, Chinese Academy of Sciences, supervised by Prof. Hui Chen from ISCAS and Ming Lu from Intel Labs China. I received my B.S. in Computer Science from University of Science and Technology, Beijing in 2023 and obtained Beijing Distinguished Graduate Award and Beijing Outstanding Graduation Thesis.

I serve as a reviewer for international conferences including ICLR, ICME and ISMAR.

My research focuses on the following areas:

  • MLLM (Qwen-VL/LLaVA/SFT/RL): general agent, R1 reasoning, and image/video understanding;
  • 3D Vision (3DGS/2DGS): digital humans, rendering, and reconstruction;
  • AIGC (Diffusion): image/animation generation;

Email  /  Github  /  Google Scholar

profile photo
Research
Mixed
[arXiv Preprint, 2025] CombatVLA: An Efficient Vision-Language-Action Model for Combat Tasks in 3D Action Role-Playing Games
Peng Chen*, Pi Bu*, Yingyao Wang, Xinyi Wang, Ziming Wang, Jie Guo, Yingxiu Zhao, Qi Zhu, Jun Song, Siran Yang, Jiamang Wang, Bo Zheng
Paper / Project / Code

We propose CombatVLA, the first efficient visual-language action model designed for combat tasks in 3D action role-playing games. For efficient decision making, our CombatVLA is a 3B model that processes visual inputs and outputs a sequence of actions to control the game (including keyboard and mouse operations).

Mixed
[arXiv Preprint, 2024] MixedGaussianAvatar: Realistically and Geometrically Accurate Head Avatar via Mixed 2D-3D Gaussians
Peng Chen, Xiaobao Wei, Qingpo Wuwu, Xinyi Wang, Xingyu Xiao, Ming Lu
Paper / Project / Code

We use 2DGS to maintain the surface geometry and employ 3DGS for color correction in areas where the rendering quality of 2DGS is insufficient, reconstructing a realistically and geometrically accurate 3D head avatar.

GazeGaussian
[arXiv preprint, 2024] GazeGaussian: High-Fidelity Gaze Redirection with 3D Gaussian Splatting
Xiaobao Wei, Peng Chen, Guangyu Li, Ming Lu, Hui Chen, Feng Tian
Paper / Project / Code

We propose GazeGaussian, a high-fidelity gaze redirection method that uses a two-stream 3DGS model to represent the face and eye regions separately.

GraphAvatar
[AAAI, 2025] GraphAvatar: Compact Head Avatars with GNN-Generated 3D Gaussians
Xiaobao Wei, Peng Chen, Ming Lu, Hui Chen, Feng Tian
Paper / Project / Code

We propose GraphAvatar, a compact method using Graph Neural Networks (GNN) to generate 3D Gaussians for head avatar animation, offering superior rendering performance and minimal storage requirements.

VARP
[NeurIPS Workshop, 2024] Can VLMs Play Action Role-Playing Games? Take Black Myth Wukong as a Study Case
Peng Chen*, Pi Bu*, Jun Song, Yuan Gao, Bo Zheng
Paper / Project

We propose a novel framework named the VARP agent, which directly takes game screenshots as input and generates keyboard and mouse operations to play the ARPG.

DiffusionTalker
[IEEE ICME, 2025] DiffusionTalker: Efficient and Compact Speech-Driven 3D Talking Head via Personalizer-Guided Distillation
Peng Chen, Xiaobao Wei, Ming Lu, Hui Chen, Feng Tian
Paper / Project / Code

We propose DiffusionTalker, a diffusion-based method that utilizes contrastive personalizer to generate personalized 3D facial animation and personalizer-guided distillation for acceleration and compression.

BYOC
[IEEE VR, 2024] Bring Your Own Character: A Holistic Solution for Automatic Facial Animation Generation of Customized Characters
Zechen Bai*, Peng Chen*, Xiaolan Peng, Lu Liu, Naiming Yao, Hui Chen, Feng Tian
Paper / Code

Given a target facial video as reference, bring your own character into our solution integrated with Unity3D, it automatically generates facial animation for the virtual character.

Internships
  • [04/2024 - 03/2025] Alibaba, Taotian
    Research intern for MLLM, focusing on MLLM-based VLA agents, including enhancing Qwen-VL series and developing advanced applications.
  • [11/2023 - 04/2024] AMD, Xilinx AI
    Research intern for Diffusion-based AIGC, especially focused on improving ControlNet and Stable Diffusion for image generation.
  • [07/2023 - 08/2023] Baidu, ACG
    Research intern for LLM evaluation, focusing on the automated evaluation of text-based question-answering tasks for the Wenxin large language model and reward model.
News
  • [06/2023] Beijing Outstanding Graduation Design (Thesis), 2023.
  • [06/2023] Beijing Distinguished Graduate Award, 2023.
Miscellaneous
Friends (click to expand, random order)

Last updated: Nov. 2024
Web page design credit to Jon Barron