I’m a Project Assistant Professor in Prof. Harada’s Lab, RCAST at UTokyo. My passion lies in developing the next generation of virtual humans and ensuring the security of the generated content. Prior to this, I obtained my Ph.D. degree in Zhejiang University under the supervision of Prof. Yi Yang. I also received my Bachelor’s and Master’s degrees in Computer Science from Nanjing University.
ECCV 2026
Introduces chunked watermarks and latent vector transformation for ownership verification and client-level leakage tracing in federated latent diffusion models.
CVPR 2025
Uses adversarial examples to suppress audio control in latent-diffusion talking-head generation.
NeurIPS 2024
Studies data leakage from federated diffusion models through multiple coordinated trojans.
ToMM 2024
Generates high-fidelity avatars under expressions unseen during training by using 3D facial priors.
ICCV 2023
Efficiently adapts audio-driven talking-head models to produce controllable emotional expressions.
Computer & Graphics 2018
Segments 3D shapes with a fully convolutional network designed for shape data.
Reviewer for TPAMI, CVPR, ICCV, IJCV, TVCG, TCSVT, TMM, ToMM, and KBS.
