Welcome to Fabien Pierre
23 January 2018
Fabien Pierre is a lecturer in computing at the IUT of Saint-Dié-des-Vosges, where he teaches web development. Originally from Aquitaine, he joined the Magrit research team last September.
This lecturer-researcher first obtained a Master’s degree in Fundamental Mathematics in 2011, then the Mathematics aggregation in 2012. But it was in 2013 that Fabien Pierre began to study image processing and edition after he obtained a Master’s degree in Applied Mathematics. He specialised in this subject until the choice of his thesis subject, which was centred on research into defogging and colouring of black and white images and videos.
Reduction of the Cauchy noise
His current research is centred on de-noising, a photographic processing technique that aims to eliminate noise due to an algorithm. The noise is translated by sparse, minute grains that are known as pepper and salt. These are present on the image and may be caused by certain sensors (of the radar type). Today, research is particularly concentrated on Cauchy noise, a random variable that is less classical in signal processing than Gaussian noise.
Within the Loria, his expertise in the processing and edition of images permitted him to join the Magrit team. Matching of images and segmentation allows the mechanical modelling of organs. He also wishes to develop algorithms for heightened reality.
Close links with the University of Kaiserslautern
Fabien Pierre also works in close collaboration with Professor Gabriele Steidl of the University of Kaiserslautern in Germany. In this dynamic, he helps different experts to colour black and white images. This technique is usually manual and laborious, and is tending to be replaced by the Colociel software, patented in 2016 in the Agency of Protection of Programmes. This digital technology consists in the transposition of a coloured image onto the image to colour, which is calculated by an estimation of geometric distortion.
This type of research has led Fabien Pierre to be published by MDPI, on the subject of transferring colours between two faces.
Since 2015, Fabien Pierre has extended his research to video. In contrast to images, the management of movement, as well as the calculation time on a GPU (processers of graphic cards), must be taken into account. Currently, a collaboration project is underway with ImaginColor concerning the colouring of videos.