|
I graduated from Southeast University with B.S. in GIS.
I got my Ph.D. degree from National Laboratory of Pattern Recognition, Chinese Academy of Sciences, supervised by Prof. Stan Z. Li and Prof. Zhen Lei, work closely with Associate Prof. Xiangyu Zhu, where I mainly research on face&vision, (e.g., 3DDFA, 3DDFA_V2).
From July 2021 to March 2025, I worked on GenAI (e.g., LivePortrait) in KlingTeam, Kwai Technology.
I am currently with ByteDance, focusing on multimodal GenAI (e.g., Lance).
|
Ghibli me ( |
Highlights |
|
and 
(
(
Publications |
|
Lance: Unified Multimodal Modeling by Multi-Task SynergyFengyi Fu*, Mengqi Huang*†, Shaojin Wu*, Yunsheng Jiang*, Yufei Huo, Hao Li, Yinghang Song, Fei Ding, Jianzhu Guo†§, Qian He, Zheren Fu, Zhendong Mao, Yongdong Zhang * Equal contribution, † Corresponding authors, § Project lead arxiv / paper / huggingface / homepage / code / 
We present Lance, a 3B native unified multimodal model for image and video understanding, generation, and editing. Lance is trained from scratch with a staged multi-task recipe, combining unified context modeling and decoupled capability pathways. The code and model weights are available at https://github.com/bytedance/Lance and Hugging Face. |
|
LivePortrait: Efficient Portrait Animation with Stitching and Retargeting ControlJianzhu Guo†, Dingyun Zhang, Xiaoqiang Liu, Zhizhou Zhong, Yuan Zhang, Pengfei Wan, Di Zhang arxiv / paper / online demo / homepage / code / 
We propose a video-driven portrait animation framework named LivePortrait with a focus on better generalization, controllability, and efficiency for practical usage. The generation speed remarkably reaches 12.8ms on an RTX 4090 GPU with PyTorch. The inference code and models are available at https://github.com/KlingAIResearch/LivePortrait. |
|
Decomposed Meta Batch Normalization for Fast Domain Adaptation in Face RecognitionJianzhu Guo, Xiangyu Zhu, Zhen Lei, Stan Z. Li TIFS (IEEE Transactions on Information Forensics and Security) (Accepted) paper / This paper proposes a novel training remedy by decomposing the model into the weight parameters and the BN statistics in the training phase. Based on decomposing, we design a novel framework via meta-learning, called Decomposed Meta Batch Normalization (DMBN) for fast domain adaptation in face recognition. |
|
Searching for Alignment in Face RecognitionXiaqing Xu, Qiang Meng, Yunxiao Qin, Jianzhu Guo, Chenxu Zhao, Feng Zhou, Zhen Lei AAAI 2021 arxiv / paper / poster / This paper proposes an efficient method Face Alignment Policy Search (FAPS) to search the optimal alignment template in face recognition. A well-designed benchmark is also proposed to evaluate the searched policy. |
|
Towards Fast, Accurate and Stable 3D Dense Face AlignmentJianzhu Guo, Xiangyu Zhu, Yang Yang, Zhen Lei, Yang Fan, Stan Z. Li ECCV 2020 arxiv / paper / supp / video1 / video2 / code / 
We propose a novel regression framework, named 3DDFA_V2, to make a balance among speed, accuracy and stability. Our model runs at over 50fps on a single CPU core (>200fps with ONNX acceleration) and outperforms other state-of-the-art heavy models simultaneously. Code and models are available at https://github.com/cleardusk/3DDFA_V2. |
|
Beyond 3DMM Space: Towards Fine-grained 3D Face ReconstructionXiangyu Zhu, Fan Yang, Di Huang, Chang Yu, Hao Wang, Jianzhu Guo, Zhen Lei, Stan Z. Li ECCV 2020 paper / code / This paper constructs a new dataset Fine-Grained 3D face (FG3D) with 200k samples for training, and proposes a Fine-Grained reconstruction Network (FGNet) concentrating on shape modi cation by warping the network input and output to the UV space. FG3D is available at https://github.com/XiangyuZhu-open/Beyond3DMM. |
|
Learning Meta Face Recognition in Unseen DomainsJianzhu Guo, Xiangyu Zhu, Chenxu Zhao, Dong Cao, Zhen Lei, Stan Z. Li CVPR 2020 (Oral Presentation, acceptance rate 5%) arxiv / paper / code / We propose a novel face recognition method via meta-learning named Meta Face Recognition (MFR) to learn a generalized model performing well on unseen domains. Besides, we propose two benchmarks for generalized face recognition evaluation. The proposed benchmarks will be available at https://github.com/cleardusk/MFR. |
|
Domain Balancing: Face Recognition on Long-Tailed DomainsDong Cao, Xiangyu Zhu, Xingyu Huang, Jianzhu Guo, Zhen Lei CVPR 2020 arxiv / paper / This paper proposes a Domain Frequency Instructor (DFI), a light weighted Residual Balancing Mapping (RBM) block and a Domain Balancing Margin (DBM) to improve generalization on tailed domains. |
|
Improving Face Anti-Spoofing by 3D Virtual SynthesisJianzhu Guo, Xiangyu Zhu, Jinchuan Xiao, Zhen Lei, Genxun Wan, Stan Z. Li International Conference On Biometrics, ICB 2019 (Oral Presentation, acceptance rate 11.2%) arxiv / paper / We present a method to synthesize virtual spoof data in 3D space to improve face anti-spoofing. The synthetic virtual samples can significantly boost the anti-spoofing performance when combined with a proposed data balancing strategy. |
|
Face Synthesis for Eyeglass-Robust Face RecognitionJianzhu Guo, Xiangyu Zhu, Zhen Lei, Stan Z. Li CCF Chinese Conference on Biometric Recognition (CCBR), Urumqi, China, Aug. 11-12, 2018 arxiv / paper / code / 
In the application of face recognition, eyeglasses could significantly degrade the recognition accuracy. In this paper, we propose to address this problem in a virtual synthesis manner. |
|
Dominant and Complementary Emotion Recognition From Still Images of FacesJianzhu Guo, Zhen Lei, Jun Wan, et al. IEEE Access, 6(1), 26391-26403, 2018 paper / code / External journal version. |
|
|
Multi-modality Network with Visual and Geometrical Information for Micro Emotion RecognitionJianzhu Guo, Shuai Zhou, Jinlin Wu, Jun Wan, Xiangyu Zhu, Zhen Lei, Stan Z. Li FG, Washington, DC. May 30 –June 3, 2017 paper / code / Winner of face micro-emotion competition held on FG 2017. |
|
3DMA: A Multi-modality 3D Mask Face Anti-spoofing DatabaseJinchuan Xiao, Yinhang Tang, Jianzhu Guo, Yang Yang, Xiangyu Zhu, Zhen Lei, Stan Z. Li 16th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), 2019 paper / We release a multi-modality 3D mask face anti-spoofing database named 3DMA, which contains 920 videos of 67 genuine subjects wearing 48 kinds of 3D masks, captured in visual (VIS) and near-infrared (NIR) modalities. |
|
Deep Manifold Attributed Graph Embedding with Graph Geodesic SimilarityZelin Zang, Siyuan Li, Di Wu, Jianzhu Guo, Yongjie Xu, Stan Z. Li Neurocomputing paper / We present a novel graph embedding framework named DMAGE (Deep Manifold Attributed Graph Embedding). |
Competitions |
|
Projects (Public) |
Contact (WeChat QR Code) |
|
|
|
|
|
|