Rémy Elie Joseph
Originally from a small village in the center of France. I followed engineering studies in telecommunication and Space Sciences in Bretany and England. I managed to come back to my original passion: astrophysics, when I got hired at CosmoStat laboratory et CEA, Saclay in 2012 as an intern. Then collaborations brought me to EPFL where I have been given the opportunity to start a PhD thesis on strong gravitational lensing.
I am interested in sports in general. I have practised Volley ball, football, badminton, surfing, skiing, hiking, running, gliding, climbing.
I have been practising fencing for 20 years for leisure and competition and am now a certified initiator. I also have interests in artistic and medieval fencing.
Developpement of image and signal processing techniques for strong lens inversion and cluster lensing analysis.
I am currently working on strong gravitational lensing. Strong gravitational lensing is a direct manifestation of Einstein's theory of general relativity that causes light's path in our Universe to be bent by massive objects. Therefore, massive galaxy clusters can act as lenses and focus the light from distant galaxies right up to our telescopes. The result: spectacular, magnified images of the first galaxies shaped as rings. These are called Einstein rings.
As a result of the focusing of light, Strong gravitational lensing can be seen as a natural telescope pointed towards the most distant regions of our sky and allows us to see galaxies that would have been too faint or too small to be observed even with our most advanced telescopes.
As the bending of light results from massive objects bending the very structure of our Universe, strong gravitational lensing is also a way of weighting galaxies including what we can not see: Dark Matter.
My work consists in inverting the effect of strong gravitational lensing to use it as a natural telescope and provide the community with clear images of the most distant objects in our sky. In the process, I am also able to estimate the mass distribution of the galaxies causing the lensing and thus deduce crucial information about the repartition of dark and luminous matter.
To do so, I propose a novel approach based on modern principles for image processing called sparsity. Sparsity allows accurate reconstruction of a signal, only assuming that it can be represented with a small number of significant coefficients in an adequate dictionnary. For instance, a degraded periodic signal can be easily reconstructed by keeping only its highest coefficients in Fourier domain.
So far I successfully developed a semi-automated Einstein ring finder based on PCA subtraction and am currently publishing a new technique for de-blending astronomical objects with different colours based on their sparsity in Starlet dictionary and on the differences in Spectral Energy Distribution (colours) in multi-band images.
PiCARD - PCA-based algorithm for galaxy profile subtraction in large surveys for the purpose of automated strong lenses finding(https://github.com/herjy/PiCARD).
MuSCADeT - A tool for colour separation of Multi-Band astronomical images (https://github.com/herjy/MuSCADeT).
- Deblending colorful objects: a morpho-spectral component analysis method for automated source separation, Statistical Challenges in 21st Century Cosmology (COSMO21), Chania, Greece, May 27th, 2016.
- Automated colour-based deblending of the Hubble Frontier Fields, EWASS2016, symposium 10: Probing the New Frontiers with Cluster Lenses, Athens, Greece, July 6th, 2016.
- Automated colour-based deblending of the Hubble Frontier Fields, Gravlens2016, Leiden, Netherlands, July 11-15, 2016.
-Automated colour-based deblending of the Hubble Frontier Fields, Cosmo21, Chania, Greece, May 24-27, 2016.
-Multi-band morpho-Spectral Component Analysis Deblending Technique (MuSCADeT), Mathematics of Signal Processing, Bonn, Germany, 14-19 february 2016.
|Scientific assistant||CosmoStat||CEA Saclay (France)||sept, 2013 - feb, 2014|
|Scientific assistant||Lastro||EPFL (Switzerland)||march, 2014 - aug, 2014|
|Doctoral Assistant||Lastro||EPFL||since oct, 2014|
|Engineering diploma in telecommunication||Specialised in image processing||Telecom Bretagne, France||2014|
|Master in Space Science and Engineering||Space Science and Engineering||University College London (UK)||2014|
|Joseph, R.; Courbin, F.; Starck, J.-L.
Published in Astronomy & Astrophysics, Volume 589, id.A2, 10 pp
|Multi-band morpho-Spectral Component Analysis Deblending Tool (MuSCADeT): Deblending colourful objects|
|R. Joseph, F. Courbin, R. B. Metcalf, C. Giocoli, P. Hartley, N. Jackson, F. Bellagamba, J.-P. Kneib, L. Koopmans, G. Lemson, M. Meneghetti, G. Meylan, M. Petkova, and S. Pires
Published in Astronomy and Astrophysics, 66, A63 March 2014
|A PCA-based automated finder for galaxy-scale strong lenses|
Paraficz, D.; Courbin, F.; Tramacere, A.; Joseph, R.; Metcalf, R. B.; Kneib, J.-P.; Dubath, P.; Droz, D.; Filleul, F.; Ringeisen, D.; SchÃ¤fer, C.
Accepted in Astronomy and Astrophysics: 21 April 2016
|The PCA Lens-Finder: application to CFHTLS|