I am a Researcher at the Shanghai Academy of AI for Science (SAIS), where I work with Prof. Siyu Zhu. My research centers on multimodal large models, including vision-language models, multimodal reasoning, multimodal generation, and foundation models. I also work on efficient inference techniques, including speculative decoding, prefix caching, and KV-cache optimization, to enable scalable deployment of large multimodal models.

I earned my Ph.D. from Zhejiang University in June 2025. Prior to SAIS, I was a Senior Vision Algorithm Engineer at Honor, leading AI-native imaging pipelines. I was also a member of the inaugural cohort of the Pujiang National Laboratory Joint Training Ph.D. Program (2022–2025), jointly advised by Dr. Yueting Chen and Prof. Tianfan Xue.

I obtained my B.S. from Central South University, where I received the National Scholarship and was recognized as an Outstanding Graduate of Hunan Province. My Master's research at Zhejiang University was advised by Prof. Huajun Feng.

I have collaborated with Siyu Zhu, Shi Guo, Zhihai Xu, Jinwei Gu, Yihao Liu, Xuelong Li, Qi Li, and Bin Zhao.

• 01.2025A paper is published in IEEE TGRS 2025 on physics-guided jitter-aware image restoration.
• 12.2024A paper is published in SIGGRAPH Asia 2024 on event-based low-light frame interpolation.
• 07.2024A paper is published in AAAI 2024 on deep restoration for linear-array imaging.
• 10.2023A paper is published in Optics Express 2023 on computational imaging.
• 08.2023A paper is published in IEEE TGRS 2023 on real-world imaging pipelines and restoration.
• 06.2023A paper is published in AAAI 2023 in computer vision and imaging.
• 11.2022A paper is published in Optics Express 2022 on deep reconstruction for interferometric imaging.

Full list on Google Scholar.

Jitter-Aware Restoration with Equivalent Jitter Model

IEEE TGRS, 2025

Physics-guided restoration for push-broom imaging.

#remote-sensing #deblurring

Paper Code

Real-world Deep Local Motion Deblurring

AAAI, 2023

Spatially-varying blur modeling for dynamic scenes.

#deblurring #real-world

Paper Code

Event-based Low-light Frame Interpolation

SIGGRAPH Asia, 2024

Sim-to-real framework for event-based low-light video.

#event-cameras #low-light

Paper Code

ASF-Transformer for Atmospheric Turbulence

Optics Express, 2023

Alternating spatial-frequency learning for turbulence compensation.

#atmospheric #imaging

Paper Code

• Email: naturezhanghn@zju.edu.cn
• GitHub: github.com/naturezhanghn
• Google Scholar: Google Scholar