Xuanzhi Liu 刘炫志
I'm currently a Research Assistant at CSCE, Shenzhen University of Advanced Technology, advised by Prof. Ruize Han and Prof. Song Wang.
Prior to that, I received my Master's degree in Information Technology from the University of New South Wales (UNSW) in June 2025. During my time at UNSW, I conducted research under the supervision of Dr. Zhengyi Yang and Dr. Junbum Kwon.
Earlier, I interned at the Research Center for Machine Vision at SIAT, Chinese Academy of Sciences, where I was advised by Prof. Zhan Song.
Recent News
- [2025/8] I start my new position @SUAT as a Research Assistant
- [2025/6] I obtain my Master's Degree in Information Technology from UNSW 🎉🎉
Research Interests
- Vision-Language Model, Computer Vision, Machine Vision
Education
University of New South Wales (UNSW), Australia
M.IT. in Artificial Intelligence
Sep. 2023 - June 2025
Guangdong University of Finance & Economics (GDUFE), China
B.Eng. in Computer Science and Technology
Sep. 2019 - June 2023
Experiences
Shenzhen University of Advanced Technology
Research Assistant
Aug. 2025 - Present
Supervisors: Prof. Ruize Han & Prof. Song Wang
UNSW Business School
Full-Stack Developer
April 2025 - May 2025, 2 months
Supervisor: Dr. Junbum Kwon
Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences
Visiting Student
July 2022 - April 2023, 10 months
Dec 2024 - Feb 2025, 3 months
Supervisor: Prof. Zhan Song
Publications & Projects
AI Comment Moderation - RAG and Classification Modelling
High-Distinction Graduation Project
This project introduces an AI-based comment moderation system that combines Retrieval-Augmented Generation (RAG) and GPT-based classification. By retrieving similar historical comments from a local vector database (ChromaDB), the system enhances classification without requiring large labeled datasets. It features a modular architecture with FastAPI backend, React frontend, and an AI Agent module for advanced multi-step reasoning. Evaluation results show strong performance (F1 score 87.3%), usability, and reliability.
A Food Package Recognition and Sorting System Based on Structured Light and Deep Learning
JCRAI 2023 Oral · Outstanding Undergraduate Thesis · Patent
We designed a sorting system for food packaging using deep learning and structured light 3D reconstruction. A pre-trained MASK R-CNN model was used to recognize the object class and structured light estimated 3D coordinates for robotic arm guidance.
© Xuanzhi Liu. Last update: .