NTU-CEE Distinguished Seminar Series: Professor JIA Haiying

Organized By

CEE Seminar Committee

Host By

Assistant Professor Yan Ran

Topic

About the Seminar

The seminar presents recent advances in the use of satellite imagery for maritime object segmentation, a task critical to environmental monitoring, autonomous navigation, and port surveillance and security. The seminar introduces a new maritime image segmentation framework that combines the Segment Anything Model (SAM) with Object-Contextual Representations (OCR), referred to as SAM-OCR, for pixellevel classification of complex maritime objects. The proposed SAM-OCR model demonstrates significant performance improvements over existing methods in the literature, particularly for challenging and previously underperforming categories such as secondary vessels within crowded scenes.

The seminar also presents a novel method for estimating material volume from a single optical image. This approach leverages material properties that relate the base dimensions of a pile to its height through the critical angle of repose. The pile geometry is mathematically modelled using a fixed-height assumption, reflecting the operational constraints of unloading cranes, whose arms rise to a limited height and move primarily in the horizontal plane. Closed-form formulas are derived to accurately estimate the volume of regular piles formed under rectilinear unloading motion.

About the Speaker

Dr. Haiying Jia is a Professor of Quantitative Business Economics at the Norwegian School of Economics (NHH). She holds a PhD in Finance from Bayes Business School and has prior industry experience in London as a Quantitative and Investment Analyst. Dr. Jia currently serves as an Honorary Professor at Bayes Business School, the United Kingdom and a Visiting Professor at Dalian Maritime University, China. She was previously a visiting scholar at the Massachusetts Institute of Technology (2018–2019) and the University of Tokyo (2022).

Her research spans shipping economics, quantitative finance, and international trade, with a strong emphasis on data-driven modelling and applied analytics. Dr. Jia has led multiple industry-collaborative research projects and published extensively in leading peer-reviewed journals. Her work contributes to both academic research and real-world applications in maritime systems, finance, and global trade.