ContextThis PhD project is part of the SPARTECHUS ANR project (2021-2025) which brings together an academic partner (CREATIS) and an ultrasound manufacturing company (IMASONIC). The project concerns two principal aspects. The first one is linked to the development of a new technology of sparse ultrasound transducers for 3D imaging (tasks lead by IMASONIC). The second concerns the development of specific driving sequences and image reconstruction algorithms as well as their joint optimization (lead by CREATIS).
ObjectiveThe specific objective of this project is to design, evaluate and implement original imaging sequences and reconstruction algorithms for 3D ultrasound imaging with sparse arrays.
MethodologyMatrix arrays were the first solution to be implemented to perform 3D ultrasound imaging. However, they require a large number of channels in the acquisition system which is rarely available in practice. Consequently, sparse arrays have been proposed. They have the advantage of permitting an individual control of each element with the current available systems, hence a full 3D steering capacity while maintaining a limited number of signals and quantity of data to process. But the sparse arrays are known to present limited signal to noise ratio compared to full matrix arrays. Their second important drawback is linked to their limited spatial sampling which is typically less than the lambda/2 spacing usually employed to avoid grating lobes. Solutions to reduce these effects on the final image are needed.
One key aspect is the layout of the probe. Whether deterministic or random positioning of the elements is the most suited approach will have to be evaluated.Then strategies to overcome the limited SNR need to be employed. In particular coded excitations are offering a vast number of possible solutions. This kind of excitations has gained much interest in the community recently thanks to the possibilities offered in terms of excitation signal by the most recent research scanners. The candidate will first have to identify the most relevant approaches in the literature, implement them and evaluate their performances. He will also probably need to propose an original scheme adapted to the specificity of the probe layout and compare it with the state of the art.
Finally advanced image reconstruction algorithms from a sparse ensemble of measurements need to be studied along with the two previous aspects. The candidate will rely on the inverse problem literature and computational imaging methodologies.
Ideally all the building blocks of the whole setup should be co-designed, taking also into account the specificities of the new technology of elements developed by IMASONIC. This makes the whole project a very complex and ambitious optimization problem.
Both simulations and experimental evaluations will have a central role and we expect to have a strong innovation impact within the duration of the PhD thanks to performing those two types of experiments simultaneously. Such experiments will be performed with a research ultrasound system available on the PiLoT platform.
ApplicationCandidates with a strong background in signal/image processing methodologies with possible previous experience in ultrasound imaging are probably the most suited. But we will evaluate all candidates having a masters’ degree with a profile fitting the subject.Candidates should send their motivation letter, CV along with the grades obtained during the last two years to email@example.com and firstname.lastname@example.org .