Haowei Kuang(匡浩玮)
Visual Intelligence Researcher
Hi, I'm Haowei Kuang(匡浩玮). I'm currently a Ph.D student at School of Intelligence Science and Technology, Peking University. My research focuses on image/video compression and enhancement.
I’m honored to study under the supervision of Prof. Zongming Guo and Jiaying Liu. Previously, I was a research assistant with Prof. Jiang Liu at SUSTech, where I earned my Bachelor's degree in Computer Science under Prof. Shiqi Yu.
Selected work · 2021—2026
Publications
Click a card to open its paper · * Equal contribution
Syntax-Driven Multi-Realism Image Compression with Consistency Guided Diffusion Model
Haowei Kuang, Wenhan Yang, Zongming Guo, Jiaying Liu
基于世界模型因果推断的图像压缩方法
匡浩玮, 杨文瀚, 郭宗明, 刘家瑛
Cross-Granularity Online Optimization with Masked Compensated Information for Learned Image Compression
Haowei Kuang, Wenhan Yang, Zongming Guo, Jiaying Liu
Consistency Guided Diffusion Model with Neural Syntax for Perceptual Image Compression
Haowei Kuang, Yiyang Ma, Wenhan Yang, Zongming Guo, Jiaying Liu
Connecting Generation with Compression: Unified Generative Coding with Semantic Guidance
Haowei Kuang, Yu Cao, Wenhan Yang, Zongming Guo, Jiaying Liu
Flash Compensated Low-Light Enhancement via Hierarchical Network Prediction
Haowei Kuang, Haofeng Huang, Wenhan Yang, Jiaying Liu
Prompt-Based Modality Bridging for Unified Text-to-Face Generation and Manipulation
Yiyang Ma*, Haowei Kuang*, Huan Yang, Jianlong Fu, Jiaying Liu
The Role of Spatial Alignment in Multimodal Medical Image Fusion Using Deep Learning for Diagnostic Problems
Xingyue Wang, Kuang Shu, Haowei Kuang, Shiwei Luo, Richu Jin, Jiang Liu
Research impact · Beyond papers
Standards, Patents & Projects
From research prototypes to standards and real-world programs
Published 2024
IEEE Standard for Neural Network-Based Image Coding
Contributed 6 proposals, with 4 integrated into IEEE Std 1857.11-2024. The standard received the 2025 IEEE SA Emerging Technology Award.
View standard ↗Large generative models
IEEE P1857.11.1
Contributed 8 proposals to the next standard for neural network-based image coding with large generative models, with 3 accepted.
Methods span low-light enhancement, controllable-fidelity compression, perceptual evaluation and generative coding.
Work on vehicle-road-cloud feature coding and multimodal understanding and reasoning for world models.
Selected achievement details+
Granted patents
- 一种使用长曝光补偿的低光照图像增强方法 CN115240022B · 2025
- 一种可调控保真度的图像压缩后处理方法 CN119693228B · 2025
Current projects
- 面向车路云一体化的高效特征数据编码与敏捷感知关键技术研究 Z241100003524010
- 面向世界模型的多模态理解与推理技术研究 Z251100008425023
Research journey
Experience & Education
A path from medical imaging to neural and generative compression
Ph.D. Student · Peking University
School of Intelligence Science and Technology. Research on neural image compression, generative coding and model-knowledge collaboration, advised by Prof. Zongming Guo and Prof. Jiaying Liu.
Southern University of Science and Technology
Worked with Prof. Jiang Liu on visual intelligence research, including multimodal medical image fusion.
B.Sc. in Computer Science · SUSTech
Studied under Prof. Shiqi Yu. Early research led to a 2021 conference paper on spatial alignment in multimodal medical imaging.