Jun Luo (罗骏)

Hello! I’m Jun Luo, a Ph.D. student at Peking University’s School of Intelligence Science and Technology, advised by Prof. Gang Zeng. My research interests lie in multimodal AI, with a focus on video generation and VLMs.

Feel free to reach out by email or WeChat. I’m always happy to connect and exchange ideas.

Research

HappyWorld-Bench

…, Jun Luo, …

Tech report

As a core contributor to HappyWorld-Bench, I developed benchmark metrics and evaluated models for the spatial (3D world model) track.

SceneAssistant: A Visual Feedback Agent for Open-Vocabulary 3D Scene Generation

Jun Luo, Jiaxiang Tang, Ruijie Lu, Gang Zeng

Code

We introduce SceneAssistant, an agent that uses vision-language models and iterative visual feedback to generate and refine open-vocabulary 3D scenes from natural language descriptions.

Zero Shot Domain Adaptive Semantic Segmentation by Synthetic Data Generation and Progressive Adaptation

Jun Luo, Zijing Zhao, Yang Liu

Code

IROS 2025

We use a diffusion model to generate two synthetic domains and a progressive adaptation strategy for zero-shot domain adaptation in semantic segmentation.

Services

Reviewer: IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)

Misc

I enjoy meeting new people and chatting about research, everyday life, and shared hobbies. Visit this page (in Chinese) to learn more about my interests beyond research.