By Gabriele Steidl
With great sadness we learned that our colleague and friend Mila Nikolova, Research Director at Centre National de la Recherche Scientique (CNRS), Microsoft Fellow, Simons Fellow, and IEEE Senior Member, is no longer with us. She passed away much too early on June 20, 2018.
Mila began her career as an engineer and science journalist in Sofia, Bulgaria, before obtaining her Ph.D. from the University Paris-Sud in Automatics and Signal Processing in 1995. She started her academic career at the Information Processing and Communication Laboratory (LTCI) at CNRS-ENST Paris, and subsequently became part of the Center for Mathematical Studies and their Applications at Ecole Normale Supérieure (CMLA-ENS), Cachan in 2003, where she became a key figure on the imaging team. She was promoted to a research director at CNRS in 2009 and had been an IEEE Senior Member since 2008.
Mila Nikolova (1962-2018). Photo courtesy of Isaac Newton Institute.
Mila was very active and influential in the international image processing community, particularly in her service to SIAM. She was an associate editor for both the SIAM Journal on Imaging Science and the SIAM Journal of Numerical Analysis, and worked diligently on the program committee for several SIAM conferences on imaging science, including the latest one this past June in Bologna, where she was already too weak to participate.
Mila's contributions to mathematics and signal processing were substantial and lasting. She was awarded with the Michel Monpetit Prize from the French Academy of Science in 2010 "for the originality and depth of her research in mathematical image processing and in solving certain inverse problems."
Her eye for interesting theoretical results that are useful in practice cannot be understated. Her work on inverse problems and models later used for compressive sensing was significant, and her study of variational models with nonsmooth data-fidelity terms, demonstrating their ability to exactly match given data, makes them very useful in practice.
With her many collaborators, both home and abroad, she made significant contributions to mathematical image processing. This includes the development of a two-phase approach to restore images corrupted by blur and impulse noise, which is still state-of-the-art, significant contributions to the image segmentation literature, including the two-phase, piecewise constant Mumford-Shah model, and a systematic convergence rate analysis of the multiplicative and additive half-quadratic minimization algorithms pioneered by Donald Geman and George Reynolds.
I have collaborated with Mila on several topics, but best remember our work on hue preserving color image enhancement which she initiated and which we wanted to continue some day. She was extremely enthusiastic in understanding human color perception and contributed many photographs for improvement which she had personally taken during her trips all over the world. She always impressed me with her understanding of fine arts and her multicultural perspective: "We are living in a 'global' world where images cross the Earth in seconds. However, the chemistry that forms color receptors in the human eye vary among the populations ...".
Mila regularly traveled abroad to present her work and exchange ideas, including numerous extended research visits, several significant ones at Hong Kong universities, among others. While her travel posed many physical and logistical challenges, this never slowed her down. Mila had many friends around the world, in particular, her French colleagues who thoughtfully took care of her during the last two months of her struggle. She will remain in her colleagues' and friends' minds through her important scientific work, her generosity, and enthusiasm.
Rest in peace my friend, with your incredible optimism and mental strength. Thank you for all the encouraging mathematical discussions and for reminding us that life is a present.
Acknowledgments: The author is grateful for some history that was translated and adapted from the Centre de Mathemematiques et Leaurs Applications tribute. She also notes the Isaac Newton Institute's thoughtful tribute.
Gabriele Steidl earned her Ph.D. and habilitation in mathematics from the University of Rostock. She is currently head of the Mathematical Image Processing Group at the University of Kaiserslautern and a consultant at the Fraunhofer Institute for Industrial Mathematics.