Jean-Michel Morel

Chair Professor in imaging at City University of Hong Kong



Research 

 Jean-Michel Morel leads a team of 25 researchers  on the mathematical analysis of image processing and the invention of new algorithms. JMM's team collaborated with French Space Agency for the design of Earth observation satellites (SPOT5, Pleiades, OTOS, SMOS-HR).  The image denoising algorithms co-invented by JMM have been implemented in more than 700 million cameraphones by DxO Labs, a world leader in software and hardware for cameras. In 2011 JMM founded Image Processing on Line (www.ipol.im), the first journal publishing reproducible algorithms in online executable articles. IPOL has collaborators in 15 universities and its public archives contain 300000 online experiments. 

JMM’s recent awards: 2010 Clay Scholar in Residence, European Research Council advanced grant 2010, Grand Prix INRIA – French Academy of Science 2013, CNRS Médaille de l'innovation 2015, Doctor Honoris Causa of U. de la Republica Montevideo 2018.

Career

Jean-Michel Morel has been in turn a PhD student at Université Paris IV, a postdoc at SISSA (Trieste, Italy), a teaching assistant at Université de Marseille Luminy, an assistant professor at Université Paris Dauphine, a professor at Ecole Normale Supérieure de Cachan and is currently a professor at City University of Hong Kong. He has also been for 20 years a permanent visiting professor at Universitat Illes Balears , and a recurrent invited professor at University of California at Los Angeles and at Duke University.

Interests

Jean-Michel Morel is a mathematician trained in nonlinear analysis with a PhD and a doctorat d’état on partial differential equations and variational methods under the guidance of Haïm Brezis. Then he developed an axiomatic theory of image analysis to derive and analyze the new PDEs and variational models arising in these fields. Working on the efficient numerical implementation of these algorithms, he has progressively become a specialist of image processing,  and has invented several algorithms now widespread in software and hardware.  He is the initiator of nonlocal methods in image processing.  Interested in image analysis as well, he has proposed a statistical theory of perception inspired from Gestalt theory and psychophysics. This theory has also found many applications for the automatic detection of objects in images and videoIn recent years, he has focused on the technological, methodological and editorial changes in applied mathematics required by the development of the web, namely the possibility to publish algorithms in online executable form.  This led him to found IPOL (image processing online, www.ipol.im.  To complete his exploration of images, Jean-Michel Morel proposes to develop mathematical models to classify the space of perceptual images. He uses computer graphics techniques to build image synthesis algorithms to explore abstract textures, shapes and images, and to develop a computational theory of decorative and abstract art.   


Industrial collaboration 

JMM and  team not only develops theories, but also accompanies their long term development and applications.The team has collaborated with the French Space Agency (CNES) on the conception of Earth observation satellites, with the major  US satellite make Planet on the image processing chain of its constellations,  with DxO labs (a world leader in image processing) on image denoising, with Huawei for the next generation mobile phones, with Agence France Presse and the French Scientific police on the detection of tampered  image and video, with Kayrros on the automatic exploitation of earth observation optical and SAR satellites to detect oil & gas activities and methane and CO2 emission sources, with Tarkett (a major floor maker)  to design natural looking textures.  

The overall average of the obtained research funds through competitive calls, grants and industrial contracts by JMM's team has amounted   to 1,100,000€ per year in the last years.