MSc thesis project proposal
Prostate cancer detection via contrast-enhanced ultrasound: modelling micro-bubble recirculation
The dynamic contrast-enhanced ultrasound is an ultrasound imaging modality that uses microbubbles to enhance the visibility of microvessels. These bubbles are smaller than the size of the blood cells and can travel through the veins without any harm. Commonly, the bubbles are injected into the arm of the patient. They traverse the body with the pressure exerted by the heart. The first pass of the bubbles from the area of interest is extracted from the signal and a model is fitted that describes the perfusion and dispersion dynamics. These features are used to classify the malignant and benign regions due to their distinctive characteristics. The time-intensity curves extracted from the recording do not only hold information about the first passage. The two-minute recording time also holds information about the recirculation of the microbubbles. The microbubbles that pass the area of interest reach back to the heart and are pumped back into the system. In this project, you will investigate estimation techniques that include the recirculation of the microbubbles. The microbubbles move through the body with the same concept that a gas fills up the room, the Brownian motion. One can model this motion using the linear density random walk model. The first pass is modeled as bubbles moving through an infinitely long tube. An assumption that can incorporate the recirculation is a circular tube.
The project will start with a literature study on the linear density random walk model, ESPRIT, and CPD-based harmonic retrieval.
• You will work on rewriting the linear density random walk model, assuming that the bubbles will reappear at the detection sites.
• You will find the poles of the system.
• Finally, you will use a CPD-based harmonic retrieval algorithm to recover the signal parameters.
A solid background in math and signal processing is required for this assignment. The student is expected to get into the theory of Brownian motion and be confident with manipulating the equations. An interest in tensor decompositions is appreciated.
dr. Borbála Hunyadi
Signal Processing Systems Group
Department of Microelectronics
Last modified: 2023-09-12