Speaker:
Assoc.Prof. Duong-Hung Pham received his Ph.D. and M.S. degrees in Applied Mathematics and Computer Science from the Grenoble Alpes University in France, in 2018 and 2014, respectively. He also earned an engineering degree from Grenoble Institute of Technology in 2013. From October 2018 to September 2019, he worked as a postdoctoral researcher in Neuroimaging with the IMAGeS Group at the Icube Laboratory, France. Since then, he has been serving as an associate professor (Maître de conférences) at the University of Toulouse (UT) and is affiliated with the IRIT – Institut de Recherche en Informatique de Toulouse (UMR CNRS 5505). 

Abstract:

This presentation introduces recent advances in Ultrasound Localization Microscopy (ULM), a promising technique designed to overcome the diffraction limit of conventional ultrasound imaging. The talk will begin with the fundamental principles of ultrafast ultrasound imaging, along with the main challenges in blood-flow estimation and super-resolution imaging.
The presentation will then focus on recent developments in post-processing methods that integrate model-based approaches and machine learning techniques. In particular, the research investigates inverse problem formulations aimed at improving the accuracy of microbubble localization and enhancing the quality of super-resolved vascular images.
Furthermore, the speaker will discuss recent efforts toward developing more efficient and interpretable algorithms for ULM reconstruction. These approaches include blind deconvolution, robust matrix decomposition, and diffusion models for ultrasound image restoration.
Finally, the talk will highlight ongoing research toward three-dimensional (3D) and transcranial ULM, with the objective of improving accuracy, computational efficiency, and clinical applicability.