À la Une

Talk Karen Egiazarian


Le Prof. Karen Egiazarian, CEO, Noiseless Imaging Oy (Finland), donnera un talk intitulé:

GPU-friendly Image Denoising

Date: Vendredi 7 décembre 2018 à 14h15

Lieu: CUI / Battelle bâtiment D, Amphithéâtre



This talk is about modelling and estimation of real camera noise and methods of noise suppression. We give a short overview of non-learning based methods, which utilize sparsity and self-similarity image priors, as well as learning based methods, in particular, convolutional neural network based ones. Next, we describe a new class of hybrid image denoisers: NN3D. They outperform counterparts, combine advantages of BM3D and DnCNN based denoisers, and provide superior quality of restored images.

Karen O. Egiazarian (Eguiazarian) (Senior Member IEEE, 1996) received M.Sc. in mathematics from Yerevan State University, Armenia, in 1981, the Ph.D. degree in physics and mathematics from Moscow State University, Russia, in 1986, and a Doctor of Technology from Tampere University of Technology, Finland, in 1994.

In 2015 he has received the Honorary Doctoral degree from Don State-Technical University (Rostov-Don, Russia). 

Dr. Egiazarian is a co-founder and CEO of Noiseless Imaging Oy (Ltd), Tampere University of Technology spin-off company.

He is a Professor at Signal Processing Laboratory, Tampere University of Technology, Tampere, Finland, leading ‘Computational imaging’ group, head of Signal Processing Research Community (SPRC) at TUT, and Docent in the Department of Information Technology, University of Jyväskyla, Finland. His main research interests are in the field of computational imaging, compressed sensing, efficient signal processing algorithms, image/video restoration and compression. Dr. Egiazarian has published about 700 refereed journal and conference articles, books and patents in these fields. He is a recipient of IS&T Service Award in 2014, is Editor-in-Chief of Journal of Electronic Imaging (SPIE), and member of the DSP Technical Committee of the IEEE Circuits and Systems Society.