About me

  • NeuroImage: New paper on metric learning using graph convolutions available.
  • GRAIL 2017: Workshop reviews, open access proceedings and presentation slides are now available at https://biomedic.doc.ic.ac.uk/miccai17-grail/reviews.html and https://biomedic.doc.ic.ac.uk/miccai17-grail/presentations.html.
  • I have recently joined Aimbrain as a Machine Learning Researcher.

    After graduating from the Ecole Centrale Paris, one of France’s top engineer schools, I did my PhD in the Center for Visual Computing at Ecole Centrale Paris under the supervision of Prof. Nikos Paragios. There, I worked on brain tumour analysis, with a strong focus on segmentation, clustering methods and atlas construction. My PhD was also partly funded by Intrasense, a french start up developing medical imaging softwares and we worked closely with Prof. Hugues Duffau, a world renown neurosurgeon specialised on diffuse low-grade gliomas.

    Over the course of my PhD, I visited the Surgical Planning Laboratory at Harvard Medical School for three months where I was supervised by Prof. William Wells III. I extended my work on tumour segmentation and registration for pre-operative and intra-operative brain registration for tumour resection.

    After finishing my PhD in 2013, I joined the Biomedical Image Analysis Group at Imperial College London working with Prof. Daniel Rueckert on the developing human connectome project (dHCP). The dHCP is an ERC synergy grant programme in collaboration between King’s College London, Imperial College London and Oxford University. The aim of the project is to create and study the first 4-dimensional brain connectivity map of early life.

    Under the scope of the dHCP, I developed methods for connectivity-driven brain parcellation through spectral clustering and Markov Random Field models, and explored the concept of deep learning on graphs for brain analysis.

    I also co-organised several workshops: