Visual Understanding / 3D Representation / Vision-Language

Chen Yanhui

I am an undergraduate researcher at Guangdong University of Technology. My work focuses on visual understanding and 3D representation, spanning open-vocabulary segmentation, language-driven 3DGS editing, and robust adaptation for vision-language models.

Portrait of Chen Yanhui
Portrait of Chen Yanhui
Chen Yanhui Guangdong University of Technology

Research

Core research directions

Visual Understanding

Open-vocabulary perception, semantic segmentation, and robust visual-language inference.

3D Representation

Language-driven 3D Gaussian Splatting editing and multi-view semantic correspondence.

Test-Time Adaptation

Reliability-aware adaptation for vision-language models under modality-specific shift.

Papers

Selected papers

Representative papers are listed below. Ongoing updates are maintained on this homepage and linked profiles.

Highlights

Patents, project, and awards

Patents (Authorized)

基于多提示协同与迁移学习的裂缝图像与视频分割方法及系统

First inventor.

Patents (Authorized)

基于深度重构网络与多维特征融合的结构损伤识别方法

Third inventor.

Provincial Innovation Project

基于深度学习的高延性混凝土裂缝识别及裂纹特征发展预测研究

Provincial undergraduate innovation project, Mar 2025 - Present.

Competitions & Honors

Recent awards

2025 National College Mathematical Contest in Modeling (Provincial Second Prize); National Innovation and Entrepreneurship Competition Provincial Gold Award; GDUT Outstanding Student Scholarship (2023-2024 Second Prize, 2024-2025 First Prize); GDUT Outstanding Student Cadre Scholarship; 2025 GDUT Top Youth League Secretary and Top Youth League Branch honors.

Experience

Leadership, internship, and skills

Student Leadership

启明星党建办公室 - 主任

负责院内党务组事务,熟悉入党流程、档案与党务信息管理,协同组织与沟通工作。

Internship

广州佰司特人力资源有限公司 - AI 猎头实习 (2024.08-2024.11)

参与 BD 方向工作,覆盖人工智能、NLP、能源、IT 与医疗器械岗位寻访,积累了谈判沟通与岗位分析经验。

Skills

Languages, engineering, and office tools

English CET-4/CET-6, Cantonese; Python, C/C++, ROS2, machine learning and deep learning; proficient in Office tools.