Souhail Hadgi

Computer Vision & Deep Learning @ Computer Science Laboratory of École Polytechnique

me.png

hadgisouhail@gmail.com

I am a final-year PhD student at École Polytechnique, advised by Maks Ovsjanikov. My research interests lie in Computer Vision and Deep Learning, with a focus on 3D data. In particular, I develop approaches to pre-train and transfer 3D encoders by using 2D and text foundation models to enable efficient global and local shape analysis.

Before my PhD, I earned an engineering degree from CentraleSupelec and an M.Sc. in Learning, Vision and Applied Mathematics (MVA) at ENS Paris-Saclay.

News

December 2025 Released PatchAlign3D, enabling efficient zero-shot 3D part segmentation via language-aligned patch features. Project page.
September 2025 Attended the PAISS Summer School in Grenoble.
June 2025 Presented a poster at CVPR 2025 in Nashville on our latest 3D-text alignment work.
February 2025 “Escaping Plato’s Cave: Towards the Alignment of 3D and Text Latent Spaces” accepted at CVPR 2025.
October 2024 Presented a poster at ECCV 2024 in Milan for our work on point cloud transfer learning.

Publications

  1. patchalign3d.jpg
    PatchAlign3D: Local Feature Alignment for Dense 3D Shape Understanding
    Souhail Hadgi, Bingchen Gong, Ramana Sundararaman, and 4 more authors
    ArXiv Preprint
    Project
  2. plato.jpg
    Escaping Plato’s Cave: Towards the Alignment of 3D and Text Latent Spaces
    Souhail Hadgi, Luca Moschella, Andrea Santilli, and 5 more authors
    CVPR 2025
    arXiv PDF
  3. supervise.jpg
    To Supervise or Not to Supervise: Understanding and Addressing the Key Challenges of Point Cloud Transfer Learning
    Souhail Hadgi, Lei Li, and Maks Ovsjanikov
    ECCV 2024
    arXiv PDF