Foundations Of Complex Understanding and Sensing

FOCUS Lab

复杂理解与感知基础实验室

我们研究视频目标跟踪与分割、复杂环境下的目标感知探测、2D/3D 重建，并关注模型安全、世界模型及多模态视觉语言大模型。

Research Themes

Visual Tracking & Segmentation Complex-Environment Perception 2D/3D Reconstruction Model Security & World Model Multimodal Vision-Language Models

About

Building foundations for complex understanding and sensing.

FOCUS Lab 围绕多媒体内容理解与人工智能安全开展研究，重点关注复杂真实环境中的视觉感知问题。

我们研究视频目标跟踪与分割、复杂环境下的目标感知探测和 2D/3D 重建，同时探索模型安全、世界模型的动态场景表示与预测，以及多模态视觉语言大模型。

Research

Core Research Directions

Visual Object Tracking & Segmentation

研究视频目标的稳定跟踪与分割，关注复杂条件下的跨域适应与跟踪可靠性。

Object Perception in Complex Environments

面向恶劣天气、伪装场景和小目标，研究目标检测与感知方法。

2D/3D Reconstruction

研究二维与三维视觉重建，以及场景几何和精细位姿对齐。

Model Security & World Model

研究模型在对抗扰动与分布偏移下的安全性和鲁棒性，并探索世界模型对动态场景的表示与预测。

Multimodal Vision-Language Models

探索视觉与语言信息的联合建模，以及多模态大模型的感知与理解能力。

People

Team

FOCUS Lab 的教师与学生成员。

副教授 · 博士生导师

Siyuan Yao 尧思远

中山大学网络空间安全学院教师，FOCUS Lab 负责人。研究聚焦视频目标跟踪与分割、复杂环境下的目标感知探测、2D/3D 重建、模型安全与鲁棒性，以及多模态视觉语言大模型。

Faculty Profile ↗ · yaosy29@mail.sysu.edu.cn

Students

Hai Long

PhD Student

Ruiqi Yu

PhD Student

Ziqi Wang

Master's Student

Shiyi Fu

Master's Student

Bingming Zhou

Undergraduate Student

News

Latest Updates

2026.08.16 · Marrying Feed-Forward 3D Reconstruction with Fine-Grained UAV Pose Alignment 在 OpenReview 发布（ACM MM UAVM 2026 Workshop Poster）。

2026.08.03 · DyFrDet 论文与代码公开，工作获 ACM MM 2026 录用。

2026.06.17 · 两篇论文获 ECCV 2026 录用：SFDATrack、SFDNet。

2026.04.09 · CAMotion: A High-Quality Benchmark for Camouflaged Moving Object Detection in the Wild 基准发布，提供野外伪装运动目标检测数据与评测。

2026 · 招生咨询：欢迎关注上述研究方向的同学联系 phdenroll123@163.com。

Publications

Selected Publications

2026

SFDATrack: Generalized Source-Free Domain Adaptive Tracking Under Adverse Weather Conditions

Siyuan Yao, Ziqi Wang, Junqi Huang, Ruiqi Yu, Wenqi Ren, Xiaochun Cao

ECCV 2026 Page Code

Adaptive Spectrum-Aware Feature Disentangled Network for Small Object Detection

Yang Guo, Zihan Yang, Feifei Kou, Yulan Hu, Ran Zhang, Siyuan Yao

ECCV 2026 Page Paper Code

DyFrDet: Towards Accurate Small Object Detection via Dynamic Frequency Suppression with Label Disambiguation

Zihan Yang, Yang Guo, Hongxing Zhang, Dan Lu, Siyuan Yao

ACM MM 2026 Paper Code

Marrying Feed-Forward 3D Reconstruction with Fine-Grained UAV Pose Alignment

Ruiqi Yu, Ziqi Wang, Haifeng Xia, Li Shengjie, Siyuan Yao

ACMMM-UAVM 2026 Poster Page

CAMotion: A High-Quality Benchmark for Camouflaged Moving Object Detection in the Wild

Siyuan Yao, Hao Sun, Ruiqi Yu, Xiwei Jiang, Wenqi Ren, Xiaochun Cao

arXiv preprint, 2026 Paper Project Code

UAGLNet: Uncertainty-Aggregated Global-Local Fusion Network with Cooperative CNN-Transformer for Building Extraction

Siyuan Yao, Dongxiu Liu, Taotao Li, Shengjie Li, Wenqi Ren, Xiaochun Cao

IEEE Transactions on Geoscience and Remote Sensing, 2026 ESI 高被引论文 Paper Code

2025

UMDATrack: Unified Multi-Domain Adaptive Tracking Under Adverse Weather Conditions

Siyuan Yao, Rui Zhu, Ziqi Wang, Wenqi Ren, Yanyang Yan, Xiaochun Cao

ICCV 2025 Page Code

UncTrack: Reliable Visual Object Tracking With Uncertainty-Aware Prototype Memory Network

Siyuan Yao, Yang Guo, Yanyang Yan, Wenqi Ren, Xiaochun Cao

IEEE Transactions on Image Processing, 2025 Paper Code

LVPTrack: High Performance Domain Adaptive UAV Tracking with Label Aligned Visual Prompt Tuning

Hongjing Wu, Siyuan Yao, Feng Huang, Shu Wang, Linchao Zhang, Zhuoran Zheng, Wenqi Ren

AAAI 2025 Page Code

CAN: Cascade Augmentations Against Noise for Image Restoration

Yanyang Yan, Siyuan Yao, Wenqi Ren, Rui Zhang, Qi Guo, Xiaochun Cao

IEEE Transactions on Image Processing, 2025 Paper

IWRN: A Robust Blind Watermarking Method for Artwork Image Copyright Protection Against Noise Attack

Feifei Kou, Yuhan Yao, Siyuan Yao, Jiahao Wang, Lei Shi, Yawen Li, Xuejing Kang

AAAI 2025 Page Code

2024

Hierarchical Graph Interaction Transformer With Dynamic Token Clustering for Camouflaged Object Detection

Siyuan Yao, Hao Sun, Tian-Zhu Xiang, Xiao Wang, Xiaochun Cao

IEEE Transactions on Image Processing, 2024 Paper Code

Join Us

We are looking for curious researchers.

欢迎对视频目标跟踪与分割、复杂环境目标感知、2D/3D 重建、模型安全与世界模型、多模态视觉语言模型感兴趣的本科生、硕士生、博士生和访问学生加入。请附上简历、成绩单、研究兴趣与代表性项目／论文链接。

Email the Lab