cv

General Information

Full Name Iris Dumeur
Languages French (native) , English (fluent), German (B1)

Education

  • 2021 to 2024
    PhD in land surface monitoring
    CESBIO, Toulouse
    • Paving the way towards foundation models exploiting Satellite Image Time Series for land surface monitoring
      • Development of three different SITS encoders - U-BARN, ALISE and MALICE
      • Self-supervised learning - masked auto encoders and Discriminative approaches
      • Temporal attention mechanism for SITS
    • Formations
      • Gretsi summer school 2022, "Methodological and societal challenges of deep learning"
      • IEEE GRSS Image Analysis and Data Fusion (IADF) summer school 2022
  • 2021
    French national engineer diploma equivalent to Master's degree
    Mines Nancy
    • Major in computer science
    • Image Processing,Advanced Data Analysis on Spatial and Temporal data,Data Science, Statistics

Experience

  • 2025-Present
    Postdoc - Change detection on mulitmodal Satellite Image Time Series
    Centre Borelli, Paris France
    • Multimodal SITS fusion - Sentinel-1, Sentinel-2, Landsat
    • Unsupervised change detection
  • 2021
    Reseach intern on bi-temporal land cover change detection using digital elevation model on stereoscopic VHR Image
    CNES, (French space Agency)
    • Bi-temporal land cover change detection using digital elevation model on stereoscopic VHR Image
    • Multi-task learning, DEM with CARS
  • 2020
    Research intern deep learning and remote sensing
    CSIRO data61, Melbourne
    • Sentinel 1 and 2 data fusion with Generative Adversarial Network (GAN)
    • Continuation of the internship as the final year school project. Assessment of the cGAN ability to simulate Sentinel 2 data and map changes
  • Sep 2019-Feb 2020
    Computer vision intern
    Preligens, ex-earthcube (PARIS)
    • Ensembling method applied to deep learning (VGG, ResNet) & machine learning (Random Forest, Gradient Boosting Tree) algorithms for ship classification in SAR image

Conferences

Open Source Projects

Skills

  • Python
    • Pytorch, pytorch-lightning, hydra, einops
  • Emacs
    • PhD and article written with orgmode
  • HPC computing
    • slurm
    • Jean-Zay and CNES HPC

Academic Interests

  • Foundation model in remote sensing
    • Multimodal data fusion
    • Self-supervised learning strategy
  • Remote sensing
    • Land surface monitoring
    • Change detection