Yann GAVET

  • Skills

    image processing, linux, augmented reality
    C/C++, python, matlab, OpenCL

  • Research Topics

    Modelisation and simulation of surfaces, mosaics and textures
    Voronoi diagrams are classically used to model mosaic patterns (like grains or cells). In the CorImMo 3D project (ANR TecSan 2012), the study of random fields allows us to randmly determine germs as well as distances to generate a spatial partition of the 2D or 3D space. The concepts and tools are based on:
    • random fields theory (covariance function, statistical moments including the skewness and kurtosis)
    • integral geometry (e.g., Minkowski functionals: area, perimeter and connectivity number for 2D fields; volume, surface area, mean breadth and connectivity number for 3D fields)
    • and mathematical morphology (e.g., openings and closures, skeletons, watersheds, granulometry).
    Mosaics
    The corneal endothelium is constituted of non-regenerative cells. We have proposed:
    • Segmentation methods based on human visual properties (Gestalt and Marr theories)
    • Segmentation evaluation tool, called epsilon dissimilarity criterion, tolerant to visual and spatial variations
    • 3D Reconstruction method (shape from focus), applied to corneal endothelium.
    CoLIP: Color Logarithmic Image Processing
    The Color Logarithmic Image Processing (CoLIP) is a recently introduced (2011) extension of the LIP mathematical framework based on abstract linear mathematics which provides a set of specific algebraic and functional operations that can be applied to the processing of color images valued in a bounded range. The CoLIP approach has been proved to be psycho-physically consistent with several laws and characteristics of the human color perception.
    Successful applications have been reported on the white balance correction in color images.

  • Teaching Areas

    Introduction to C programming, Signal Processing (L3)
    Introduction to image processing (M1), advanced image processing (M2), High Performance Computing (openmp, openCL).
    Taught in French and/or English

  • Biography

    His research is at the crossroad of mathematics (for the scientific background), physics (for the image formation rules), computer engineering (for testing and constructing the practical solutions) and the application fields.

  • Education

    Engineer, 2001
    Master thesis in image processing, 2004
    PhD thesis in image processing, 2008
    Habilitation 2015

  • Career History

    Assistant Professor
    Leader of the CorImMo3D project (ANR 2012, 650 K€)

    4 months (2017) detachment in Thalès-Angénieux, Saint-Heand, France.
    formerly: IT developer for CNES (Space French Agency)

53 documents

  • Nima Hatami, Yann Gavet, Johan Debayle. Classification of Time-Series Images Using Deep Convolutional Neural Networks. Proceedings of SPIE, the International Society for Optical Engineering, SPIE, The International Society for Optical Engineering, 2018, 10696, pp.UNSP 106960Y. ⟨10.1117/12.2309486⟩. ⟨hal-01869027⟩
  • Hamza Moussaoui, Jérôme Laurencin, Yann Gavet, Gérard Delette, Maxime Hubert, et al.. Stochastic geometrical modeling of solid oxide cells electrodes validated on 3D reconstructions. Computational Materials Science, Elsevier, 2018, 143, pp.262-276. ⟨10.1016/j.commatsci.2017.11.015⟩. ⟨hal-01655845⟩
  • Solmaz Boroomandi Barati, Nicolas Pionnier, Jean-Charles Pinoli, Stéphane Valette, Yann Gavet. Investigation spatial distribution of droplets and the percentage of surface coverage during dropwise condensation. International Journal of Thermal Sciences, Elsevier, 2018, 124, pp.356 à 365. ⟨10.1016/j.ijthermalsci.2017.10.020⟩. ⟨hal-01774493⟩
  • Martin Hélène, Solmaz Boroomandi Barati, Nicolas Pionnier, Jean-Charles Pinoli, Valette Stéphane, et al.. Segmentation of gray scale images of dropwise condensation on textured surfaces. International Journal of Circuits, Systems and Signal Processing, North Atlantic University Union, In press, 12, pp.339 à 348. ⟨hal-01919137⟩
  • Nima Hatami, Yann Gavet, Johan Debayle. Classification of Time-Series Images Using Deep Convolutional Neural Networks. 2017 The 10th International Conference on Machine Vision (ICMV 2017), ICMV Committees, Nov 2017, Vienne, Austria. ⟨10.1117/12.2309486⟩. ⟨hal-01743695⟩
  • Solmaz Boroomandi Barati, Jean-Charles Pinoli, Valette Stéphane, Yann Gavet. Random and geometric model of droplets nucleation on textured surfaces. Séminaire LABEX MANUTECH-SISE, EQUIPEX MANUTECH-USD; LABEX MANUTECH-SISE, Oct 2017, Marcoux, France. ⟨hal-01924135⟩
  • Helene Martin, Solmaz Boroomandi Barati, Jean-Charles Pinoli, Stephane Valette, Yann Gavet. Recognition the droplets in gray scale images of dropwise condensation on pillared surfaces. 7th International Conference on Computer Science, Engineering & Applications (ICCSEA 2017), Academy & Industry Research Collaboration Center (AIRCC), Sep 2017, Copenhague, Denmark. pp.115 à 126, ⟨10.5121/csit.2017.71109⟩. ⟨hal-01634524⟩
  • Clothile Jumelle, Garcin Thibaud, Anne-Sophie Gauthier, Glasson Yäel, Aurelien Bernard, et al.. Considering 3D topography of endothelial folds to improve cell count of organ cultured corneas. Cell and Tissue Banking, Springer Verlag, 2017, 18 (2), pp.185 à 191. ⟨10.1007/s10561-017-9624-7⟩. ⟨hal-01609175⟩
  • Klervi Rannou, Gilles Thuret, Phillipe Gain, Jean-Charles Pinoli, Yann Gavet. A novel quantitative methodology for age evaluation of the human corneal endothelium. QCAV 2017 - 13th International Conference on Quality Control by Artificial Vision, Technical Committee on Industrial Application of Image Processing, Japan Society for Precision Engineering; Chuo University, May 2017, Tokyo, Japan. pp.103381D, ⟨10.1117/12.2263670⟩. ⟨hal-01928564⟩
  • Hamza Moussaoui, Johan Debayle, Yann Gavet, Gérard Delette, Maxime Hubert, et al.. 3D geometrical characterization and modelling of solid oxide cells electrodes microstructure by image analysis. Thirteenth International Conference on Quality Control by Artificial Vision 2017 (QCAV 2017), Chuo University, May 2017, Tokyo, Japan. pp.1033804, ⟨10.1117/12.2264376⟩. ⟨hal-01526466⟩