Learning Stochastic Geometry Models for Ship Identification in Synthetic Aperture Radar Images
1 - 31 July 2025
Biography
Dr Josiane Zerubia’s (Dr Eng, PhD, Habilitation) main research interest is in image processing using probabilistic and artificial intelligence models for remote sensing. She also works on parameter estimation, statistical learning and optimization techniques. She has been a permanent research scientist at Institut National de Recherche en Informatique et en Automatique (INRIA) since 1989 and Director of Research Exceptional Class since 2023. She has been heading several research teams in remote sensing at Inria in Sophia-Antipolis, France, since mid-1995: Pastis, Ariana, Ayin and most recently Ayana, which is using knowledge in stochastic modeling, image processing, artificial intelligence and remote sensing for the “New Space”. She wrote with her collaborators 4 books, 17 book chapters, 91 journal papers and 273 conference publications. Her h-index is 59, i-10 is 220, total citation number is 15 309 (on July 23, 2024). During the last few years (2021-2024), she was member-at-large of the Awards Board of the IEEE SP Society, member of the Best Paper Award Committee for EURASIP JIVP, member of the IAPR Fellow Committee, member of the Senior Editorial Board of IEEE SP Magazine and senior member of the IEEE Women in Signal Processing Committee. She is a Fellow of the IEEE (2003), the EURASIP (2019) and the IAPR (2020), and IEEE SP Society Distinguished Lecturer (2016-2017). She is also Doctor Honoris Causa of the University of Szeged in Hungary (2020).
Research Summary
Over the last few decades there has been a growing interest in Synthetic Aperture Radar (SAR) imaging on account of its importance in applications such as mapping, search and rescue, and target recognition and tracking. Modern SAR systems can produce high quality pictures of the Earth’s surface while avoiding some of the shortcomings of other forms of remote imagining systems, like limitations during nighttime and seeing through cloud cover. Its value also extends to areas like marine spatial planning and habitat restoration. In maritime applications, accurate analytics of SAR imagery is not only important in isolation, but also in the detection and characterization of ship wakes. These provide key information for tracking (illegal commercial activities) vessels and are also useful in classifying the characteristics of the wake generating vessel and hence estimate their velocity. Until recently, one of the main factors hampering research into sea surface modelling was the lack of data of sufficiently high resolution (pixels need to be typically smaller than a few meters) and accuracy. SAR technologies have however shown remarkable progress in recent years, and the availability of remotely sensed data of the Earth and sea surface is continuously growing.
On the other hand, artificial intelligence (AI) and machine learning have reached a significant level of maturity, with many methods having been developed in the field of object detection and classification in remote sensing images.
We plan in collaboration, to investigate the recent vehicle detection in optical satellite imagery, Zerubia’s group proposed harnessing CNN models information extraction abilities in combination with point process interaction models, using CNN outputs as data terms, and to incorporate it into Achim's group ship detection technology that uses SAR.
During her one month stay at University of Bristol in July 2025, Dr. Zerubia plans to give an Open Seminar on "learning point process models for object detection in remote sensing" and a Postgraduate Seminar on "fully convolutional and feedforward networks for the semantic segmentation of remotely sensed images"
Dr Zerubia's lectures and seminars will be listed on our Events page in due course.
You can contact Dr Zerubia's host Professor Alin Achim for further information.
