Internship in 3D multiparametric ultrasound segmentation
Duration: 4-6 months
Location: 2-10 rue d’Oradour-sur-Glane, 75015 Paris, France
Contact: send your CV to beatrice.walker@espci.fr
This internship is part of a project aiming to apply multiparametric ultrasound imaging to precisely characterizing tumors for breast cancer diagnosis. 3D maps of tissue stiffness, vascularization and tissue fiber structure will be used to infer tumor characteristics such as stage and grade. This requires precise localisation of the tumor boundaries, which will not be possible on standard Bmode ultrasound imaging.
This internship focuses on the development of an automated segmentation technique using the different registered quantitative 3D maps to delineate the tumor. This will be done using existing data obtained from ex-vivo breast tumor samples for which radiologist delineation is already available as a gold standard. The intern will be responsible for designing an efficient and automated segmentation algorithm using techniques ranging in complexity from gradient-based deformable contours to deep learning.