Guillaume Tochon

• Site: Paris
• Group: image processing and pattern recognition
• Axis: machine learning and applications

Short Bio

I obtained a M.Sc degree in electrical engineering, with a major in signal and image processing, from the Grenoble Institute of Technology, France, in 2012. From January to June 2012, I was also a Predoctoral Fellow with Carnegie Institution for Science, Stanford University, USA. I received a PhD degree in signal and image processing from the Université Grenoble Alpes, France, in 2015. I conducted my research at GIPSA-lab within the SigmaPhy (‘‘Signaux et Images Physiques’’) team, under the supervision of Jocelyn Chanussot and Mauro Dalla Mura. I was also a visiting scholar with the Department of Mathematics, University of Californie, Los Angeles, USA from June to August 2013. From 2015 to 2016, I was a Temporary Assistant Professor (‘‘ATER’’) with the Grenoble Institute of Technology and GIPSA-lab.
Since September 2016, I am an Assistant Professor with EPITA and I conduct my research with EPITA Research Laboratory, within the Image Processing and Pattern Recognition team.
Since September 2022, I am also an Invited Researcher at the Institut de Mécanique Céleste et de Calcul des Éphémérides (IMCCE, Observatoire de Paris), within the PEGASE team.

Research

Research interest

My research interests lie in the fields of mathematical morphology (notably the use of hierarchical representations), statistical signal processing and machine/deep learning (sometimes all interacting together), with applications in space imagery (astrometry) and satellite imagery (particularly multi/hyperspectral image processing).

Examples of on-going projects include:

• Learning structuring elements and morphological operation with deep neural networks (former work of A. Kirszenberg and R. Hermary), jointly with É. Puybareau and J. Angulo (CMM).
• Learning the temporal dynamics directly from Sentinel-2 multispectral time series data (phD work of A. Frion, ANR JCJC Lemonade), jointly with A. Aïssa El Bey (IMT-Atlantique), L. Drumetz (IMT-Atlantique) and M. Dalla Mura (Grenoble-INP).
• Detection and classification of meteors and satellites in 1 second long exposure pictures of the night sky, jointly with É. Puybareau, J. Fabrizio and J. Vaubaillon (IMCCE).
• Large scale analysis of the CASSINI image database for space object detection and classification (PhD work of G. Quaglia), jointly with V. Lainey (IMCCE)

Publications

For a full list of publications, you can check:

• Google Scholar Page
• DPLB

Supervision

I currently supervise the PhD work of:

• B. Esteban (since Nov. 2020). Topic: ‘‘Modern and efficient implementations of Mathematical Morphology algorithms for image processing and noise estimation’’. Joint supervision with D. Verna and E. Carlinet)
• A. Frion (since Oct. 2021). Topic: ‘‘Learning and modeling spectral dynamics of satellite image time series’’. Joint supervision with A. Aïssa El Bey (IMT-Atlantique), L. Drumetz (IMT-Atlantique) and M. Dalla Mura (Grenoble-INP)), in the framework of the ANR JCJC Lemonade (PI L. Drumetz)
• G. Quaglia (since Sept. 2022). Topic: ‘‘Application of IA techniques to the analysis of the Cassini ISS images’’. Joint supervision with V. Lainey (IMCCE)

Former students:

• J. Rivet, who defended her [https://www.lrde.epita.fr/wiki/Affiche-these-JR PhD] in July 2020, on ‘‘Non-iterative methods for image improvement in digital holography of the retina’’ (joint supervision with T. Géraud, M. Paques (Sorbonne Université), S. Meimon (ONERA) and M. Atlan (Institut Langevin))
• J. Estopinan (M2 intern between Feb. 2020 and Jul. 2020), on the subject ‘‘Temporal Dynamics Learning for Hyperspectral Images Unmixing’’ (joint supervision with L. Drumetz (IMT-Atlantique))
• R. Hermary (EPITA research student, Feb. 2021- Feb. 2022). Topic: ‘‘Learning structuring elements and morphological operation with deep neural networks’’ (joint supervision with É. Puybareau)
• M. El Gemayel (EPITA research student, Sept. 2020- Sept. 2021). Topic: ‘‘Detection and classification of meteors and satellites in 1 second long exposure pictures of the night sky’’ (joint supervision with É. Puybareau)
• B. Parsy (EPITA research student, Feb. 2020 - Feb. 2021), on the subject ‘‘Sensitivity of hierarchical representations to noise in images’’
• T. Buatois (EPITA research student, Feb. 2019 - Feb. 2021), on the subject ‘‘Smart and robust segmentation of medical images using neural networks’’ (joint supervision with É. Puybareau)
• A. Kirszenberg (EPITA research student, Feb. 2019 - Feb. 2020), on the subject ‘‘Learning morphological operations with deep neural networks’’ (joint supervision with É. Puybareau)
• B. Esteban (EPITA research student, Sept 2018 - Feb. 2020), on the subject ‘‘Noise estimation from the tree of shapes and non parametric statistics’’
• T. De Carvalho (EPITA research student, Sept 2018 - Feb. 2020), on the topic ‘‘Learning Model Checking’’ (supervision with E. Renault)

Teaching

I am teaching the following courses at EPITA:

• Mathematics for signal processing (ING1 level - third year of undergrad studies). Lectures: 14h / practical labs: 12h
• Data compression (ING1 level - third year of undergrad studies). Lectures: 9h / practical labs: 3h
• Introduction to image processing (ING1 level - third year of undergrad studies). Practical labl: 12h
• Convex Optimisation (ING2 IMAGE - first year of grad studies). Lectures: 18h / Practical labs: 18h
• Introduction to Machine Learning (ING2 IMAGE - first year of grad studies). Lectures: 5h / practical labs: 12h
• Constrained optimization (ING3 IMAGE - second year of grad studies). Lectures: 10h / practical labs: 6h