This year, I am taking over the LASTIG ML/CV reading group. If you are interested, you are welcome to join!

Every two weeks, there will be 1 or 2 presentations followed by a short discussion for a total duration of 1 to 1.5 hour. Each presentation will summarize an article/method/ongoing work that worth sharing with other students and researchers. We will talk about both technical and application/remote sensing papers.

The group meetings will be held every two weeks of Fridays at 2PM, starting from 23 September. Room to be defined. We expect everyone who join the reading group to regularily present, you can also invite researchers form other labs to come and present their work!



  • 31/03/2023 - Pixel-wise Agricultural Image Time Series Classification: Comparisons and a Deformable Prototype-based Approach
  • 03/03/2023 - NLP: from IGN use cases to ChatGPT
  • 10/02/2023 - From Images to 3D Models
    • Charles Villard - LASTIG/IGN - Hypergraph Structure From Motion (HyperSFM)
      Slides | Scholar
    • Lulin Zhang - LASTIG/IGN - NeRFactor: Neural Factorization of Shape and Reflectance Under an Unknown Illumination
      Slides | Scholar
  • 27/01/2023 - PhD student updates
    • Quentin Potié - LASTIG/IGN - Landmarks segmentation in maps with Deep Learning
      Slides | Scholar
    • Solenn Tual - LASTIG/IGN - A Benchmark of Nested Named Entity Recognition Approaches on Historical Documents
    • Chahine-Nicolas Zede - LASTIG/IGN - Point cloud based large-scale place recognition - Application to the prevention against fake news
    • Martin Cubaud - LASTIG/IGN - Updating land cover data: a multi-source and multi-modal approach for change qualification and land use characterization by deep learning
    • Florent Geniet - LASTIG/IGN - From pointcloud to 3D urban model
  • 13/01/2023 - Unsupervised Multi-Domains Adaptation for Semantic Segmentation of Very High Resolution Aerial Images
  • 16/12/2022 - Modeling of Dense Vegetation from Aerial LiDAR Scans
  • 2/12/2022 - Recent methods for self-supervised learning in computer vision (MAE & Contrastive Learning)
  • 18/11/2022 - Transfer Learning of Convolutional Neural Networks for Texture Synthesis and Visual Recognition in Artistic Images
  • 21/10/2022 - Diffusion models
  • 07/10/2022 - Next best view & Historical maps
    • Antoine Guédon - IMAGINE/ENPC - SCONE: Surface Coverage Optimization in Unknown Environments by Volumetric Integration (NeurIPS2022)
      Slides | Scholar
    • Dina El Zein - IMAGINE/ENPC & LASTIG/IGN - Adversarial Generation of Historical Maps
  • 23/09/2022 - Historical maps analysis
    • Sidi Wu - IKG/ETHz - Retrieving spatial-temporal dynamics of hydrology from scanned historical maps
      Slides | Scholar
    • Yizi Chen - LASTIG/IGN & LRDE/EPITA - Retrieving spatial-temporal dynamics of hydrology from scanned historical maps
      Slides | Scholar | Website