Qi Song

I'm a PhD student at Department of Computer Science, Hong Kong Baptist University, supervised by Dr. WAN, Renjie.

I obtained my bachelor's degree in Computer Science from Beijing Normal University in 2023. Since Sep. 2023, I have also been a research intern at NVIDIA AI Technology Center.

Currently, I am focusing on AI safety for generative model.

Please feel free to email me or just add my wechat (qsong2001)!

Email  /  Google Scholar  /  CV

Publication

Generating Multi-view Adversarial Examples for Visual Geometry Grounded Transformer
Qi Song, Ziyuan Luo, Haoliang Han and Renjie Wan
in ECCV ( Oral, Top 1.6%), 2026
Paper / Code

Creating Blank Canvas Against AI-enabled Image Forgery
Qi Song, Ziyuan Luo, and Renjie Wan
in AAAI, 2026
Paper / Code

Align 3D Representation and Text Embedding for 3D Content Personalization
Qi Song, Ziyuan Luo, Ka Chun Cheung, Simon See and Renjie Wan
in ACM MM (Oral), 2025
Paper / Code

Geometry Cloak: Preventing TGS-based 3D Reconstruction from Copyrighted Images
Qi Song, Ziyuan Luo, Ka Chun Cheung, Simon See and Renjie Wan
in NeurIPS, 2024
Project Page / Paper / Code

Protecting NeRFs' Copyright via Plug-and-Play Watermarking Base Model
Qi Song, Ziyuan Luo, Ka Chun Cheung, Simon See and Renjie Wan
in ECCV, 2024
Project Page / Paper / Code

Securing INR-Based Steganography with Quantum Circuit-driven Weight Initialization
Qi Song, Haoliang Han, Ziyuan Luo, Qi Jun and Renjie Wan
in ICASSP (Oral), 2026
Paper / Code

Degradation Conditioned GAN for Degradation Generalization of Face Restoration Models
Qi Song, Wu Shi, Guojing Ge and Liang Chang
in ICIP (Oral), 2023
Paper

Benchmarking Wild Bird Detection in Complex Forest Scenes
Qi Song, Yu Guan, Xi Guo, Xinhui Guo, Yufeng Chen, Hongfang Wang, Jianping Ge, Tianming Wang and Lei Bao
Ecological Informatics, 2024
CV4Animal / Paper

Taisu: A 166M Large-scale High-Quality Dataset for Chinese Vision-Language Pre-training
Yulong Liu, Guibo Zhu, Bin Zhu, Qi Song, Guojing Ge, Haoran Chen, GuanHui Qiao, Ru Peng, Lingxiang Wu and Jinqiao Wang
in NeurIPS (Spotlight), 2022
Paper / Code

Under Review

WFF3D: Watermarking Feed-Forward 3D Gaussian Splatting Models
Code

HiD-Splat: One-Pass Attribute-Space Watermarking for 3D Gaussian Splatting
Code