Hi there 
Welcome to my academic base!
I am a fourth-year undergraduate(2022-2026) at Nankai University, pursuing my Bachelor’s degree in Software Engineering. My research is currently under the guidance of Prof. Mingming Cheng and Prof. Yun Liu at the Media Computing Lab, Nankai University. I am also an incoming graduate student at the iMoon Lab, where I will be advised by Prof. Yue Gao.
Research Interests
My current research focuses on three key areas:
Image Generation and Editing: Exploring generative models such as Stable Diffusion for controllable image generation, editing, and detail restoration, with an emphasis on improving model generalization.
Few-shot Learning: Developing methods for efficient learning from limited data, focusing on techniques like metric learning, prompt tuning, and cross-modal transfer to enhance data efficiency.
Hypergraph Learning: Investigating hypergraph-based models to capture high-order relationships and improve representation power, structural consistency, and scalability in complex learning systems.
News
- I have set up a Blog Site, welcome everyone to visit!
Research Experience
Publications
Guohuan Xie, Xin He, Dingying Fan, Le Zhang, Ming-Ming Cheng, and Yun Liu†
We propose Syn4Seg, a generation-enhanced framework for generalized few-shot semantic segmentation that expands novel-class coverage with diverse synthetic images, support-guided pseudo-label enhancement, and SAM-based boundary refinement. Experiments on PASCAL-5i and COCO-20i show consistent improvements in both 1-shot and 5-shot settings while maintaining strong base-class performance.
[arXiv]
Guohuan Xie, Syed Ariff Syed Hesham, Wenya Guo, Bing Li,
Ming-Ming Cheng, Guolei Sun, and Yun Liu†
This survey comprehensively reviews Video Scene Parsing (VSP), encompassing VSS, VIS, VPS, VTS, and OVVS. It analyzes architectural advances from CNNs to Transformers, addresses core challenges like temporal consistency, and provides unified benchmarks, datasets, and future research perspectives.
[arXiv]
Projects
NKU-AI-Assistant is a conversational AI system using multi-agent architecture and RAG. It features ChainMind, MetaAgent, and NeoGraph for dialogue, reasoning, and knowledge retrieval, respectively, supported by web, video, file parsing, and translation plugins for enhanced versatility.
[code] [demo]
Immortal Verse is an RPG set in Tang-dynasty China, blending history, poetry, and AI-driven interaction. Players explore landscapes, complete quests, engage in battles, learn classical poetry, and generate poems with AI, enjoying immersive cultural experiences and intelligent dialogues centered on poet Li Bai.
[code] [demo]
Academic Service
- reviewer:IJCAI2026,ICMR2026
Awards
- 2025:Migu M-Zone Most Valuable Application Award(3/1143, Top 0.3%) – 50,000 CNY
- 2025:BYD Scholarship – 10,000 CNY
- 2025:National Scholarship(1/130, Top 0.8%) – 10,000 CNY
- 2024:National Scholarship(1/127, Top 0.8%) – 10,000 CNY
- 2024:Second Prize, National Level, China Mathematical Contest in Modeling（Top 2%）
- 2023:Gong-Neng Scholarship of Nankai University(5%) - 5000CNY
- 2023,2024,2025: Outstanding Student of Nankai University