PhD Position in Inverse Problems for Biomedical Imaging

This doctoral position focuses on developing novel non-Gaussian image denoising techniques that preserve fractal properties to improve the multi-scale analysis of X-ray mammographies. Hosted at the Laboratoire des Sciences du Numérique de Nantes (LS2N) and employed by the CNRS, the successful candidate will combine model-based and data-driven approaches, integrating optimization procedures with deep learning. The research is part of the ANR-funded UNGRAM project, which aims to better evaluate breast cancer risk by analyzing self-similarity in mammographic textures. The PhD candidate will collaborate closely with an interdisciplinary consortium of researchers, a postdoctoral fellow, and incoming master's interns.

• Develop non-Gaussian image denoising methods that preserve the fractal and self-similarity properties of mammographic textures
• Explore and combine model-based approaches and data-driven machine learning techniques for robust image processing
• Design and implement optimization procedures combined with deep learning architectures
• Analyze X-ray mammographies through multi-scale analysis to assist in breast cancer risk assessment
• Collaborate with project partners, postdoctoral researchers, and student interns within the UNGRAM consortium
• Participate in international workshops, scientific conferences, and project meetings