I am a scientific assistant at ETH Zurich’s Photogrammetry and Remote Sensing (PRS) Group. I currently work with Konrad Schindler, Jan Dirk Wegner, and several PhD students in the group. I focus my research on both fundamental research on computer vision and machine learning, as well as on environmental and humanitarian applications. In recent projects I have performed monocular depth estimation for snow height mapping from satellite images, super-resolved depth images using RGB guidance, and mapped vulnerable populations in African countries, for instance. More fundamental research topics include robust evaluation of debiasing algorithms using hypothesis testing and building implicit ensembles using feature modulation for uncertainty estimation.
I have previously received my doctoral degree from Institut Polytechnique de Paris under the supervision of Bertrand Le Saux, Alexandre Boulch, and Yann Gousseau. My thesis explored the usage of convolutional neural networks for change detection and related data quality problems, such as label noise and domain shifts.
I am interested in several topics in computer vision, especially mapping, multitemporal analysis, computational photography, and methods to support sustainable development. I also recently got especially interested in the applications of implicit neural representations outside the context of rendering.