Florent Imbert

Bio

After a master's degree in applied mathematics at the La Rochelle University. I pursued a PhD in computer science at IRISA on the SHADoc (formerly Intuidoc) team, working under the supervison of Eric Anquetil, Romain Tavenard and Yann Soullard.
I contribute to the KIHT project which consists in designing and developing the deep learning AI that reconstructs online handwriting trajectories from the DigiPen’s kinematic sensors (accelerometers, gyroscope, magnetometer, force sensors). This project is in partnership with Stabilo, KIT and Learn&Go.
Currently, I am a postdoctoral researcher at Luleå University of Technology (Sweden) in the Machine Learning team, where I am engaged in research on suitable machine learning.

News

Thesis defense:

    ‘Design of a deep neural network architecture dedicated to handwriting synthesis from kinematic sensors of a digital pen’. It was presented to the following jury:
  • Reporter:
    • Andreas Fischer - Full professor - University of Applied Sciences and Arts Western Switzerland (HES-SO)
    • Clément Chatelain - Maître de conférences HDR, INSA Rouen Normandie​
  • Examiner:
    • Elisa Fromont - Professeur des universités, Université de Rennes​
    • Nicolas Ragot - Maître de conférences HDR, Polytech Tours​
  • Thesis supervisor:
    • Eric Anquetil Professeur des universités, INSA Rennes​
    • Romain Tavenard Professeur des universités, Université Rennes 2​
    • Yann Soullard Maître de conférences, Université Rennes 2
  • Abstract: This thesis focuses on a digital pen equipped with kinematic sensors, and its aim is to reconstruct the in-line trace of handwriting. We introduce a new processing pipeline that associates pen sensor signals with the corresponding writing trajectory. Based on Dynamic Time Warping to align the signals and an architecture inspired by Temporal Convolutional Networks Additionally, we present a Mixture-Of-Experts (MOE) approach to enhance the focus and understanding of each aspect of handwriting, comprising a touching expert model for pencil touches and a pen-up expert model for pen trajectories. A significant challenge is the variation in captured signals between adults and children, due to differences in speed and confidence in handwriting gestures. We address this through a domain adaptation approach. Furthermore, we introduce a new public benchmark dataset to support future research and comparisons in the field of handwriting reconstruction.

© Florent Imbert. Some rights reserved.