Reuben Dorent

Researcher in Computer Vision

I am currently a postdoctoral fellow at Harvard Medical School working with Prof. Sandy Wells. Prior to that, I did my PhD under the supervision of Prof. Tom Vercauteren at King’s College London.

I believe that all available data, even incomplete data and data with missing, incomplete, or partial annotations should be exploited to build robust and flexible machine learning models.

I am primarily interested in computer vision with a focus on medical applications. My main research interests are:

  • Weakly Supervised Learning: Learning image segmentation and registration using weak labels (partial annotations, scribbles, points).
  • Domain Adaptation: Transfering the knowledge learned from one modality to another.
  • Representation Learning: Discovering joint representations of multimodal data to handle incomplete sets of input data at training and inference times.

Contact: r[surname]

Selected publications

  1. CrossMoDA 2021 challenge: Benchmark of cross-modality domain adaptation techniques for vestibular schwannoma and cochlea segmentation
    Reuben Dorent*, Aaron Kujawa, Marina Ivory, and 37 more authors
    Medical Image Analysis, Oct 2022
  2. Unified Brain MR-Ultrasound Synthesis Using Multi-modal Hierarchical Representations
    Reuben Dorent*, Nazim Haouchine, Fryderyk Kogl, and 9 more authors
    In Medical Image Computing and Computer Assisted Intervention – MICCAI 2023, Sep 2023
  3. Learning joint segmentation of tissues and brain lesions from task-specific hetero-modal domain-shifted datasets
    Reuben Dorent*, Thomas Booth, Wenqi Li, and 5 more authors
    Medical Image Analysis, Jan 2022