Emmanuel Chevallier: PhD at the CMM, MINES ParisTech

I am Emmanuel Chevallier from France. I finished my master degree from the university of Rennes 1 and my engineering diploma from Telecom Bretagne in 2012. Since then, I am a PhD student in the Center for Mathematical Morphology (CMM), MINES ParisTech, under the supervision of Jesus Angulo.

Since its apparition, mathematical morphology has evolved with the development of computer science. At first image processing was limited to binary images. After the apparition of gray levels, images became color then multi-spectral. The theory of image processing had to adapt to each of these evolutions. My Ph.D. originally aimed at extending the theory of mathematical morphology to non standard images such as tensor images (e.g., DTI used in medical imaging). This work is the direct continuation of my predecessor.

The Ph.D. takes mainly place at the CMM in Fontainebleau however I had the opportunity to spend 3 months at the University of Groningen.

My work has now two distinct orientations. The first part concerns the theoretical development of mathematical morphology to “non-grayscale” images. My current contributions consist in the formalization of the discontinuity issue of total orders in metric spaces, and in a proposition of morphological operators for images valued in a set of independent labels.

The second part concerns the density estimation on Riemannian manifolds applied to image and signal processing (e.g., computation of histogram of images valued in Riemannian manifolds). The second problem originally stemmed from the first one. However it has now become a full-fledged subject of my Ph.D. I have mainly worked on symmetric spaces such as the hyperbolic space, the space of symmetric positive definite matrices and on the manifold of colours endowed with perceptual metrics.

The taste for research pushes me to find an academic post-doc after my Ph.D. Here is my personal webpage: http://cmm.ensmp.fr/~chevallier/