Hans Knutsson has more than 400 publications in image processing and machine learning. The 10 most cited have a total of more than 4000 citations, h-index = 40. He is one of the founding members of CMIV - the Center for Image Science and Visualization. He has been, or is, the supervisor of 21 PhD students.
He obtained his PhD in 1982. In 1984 - 1986 he held a Postdoctoral Fellowship position at the Rockefeller University Neurobiology Laboratory headed by Nobel Laureate Torsten Wiesel.
He suggested and developed a novel method of representation and estimation of local structure of multi-dimensional signals using tensors and tensor fields. A result of this research was a method for handling irregularly sampled and uncertain data termed 'Normalized Convolution'. Advanced spatio-temporal filtering, registration and segmentation techniques continues to be the main targets of his research. In magnetic resonance imaging (MRI) he introduced Canonical Correlation based fMRI analysis. More recent important results are: MRI-reconstruction of beating heart-sequences, real-time fMRI brain-computer interface and 4D denoising of medical image data.
In collaboration with LMI , the Laboratory for Mathematics in Imaging at Harvard Medical School and the Surgical Planning Laboratory in Boston he has also started a new line of research aiming at developing a common theoretical framework for sampling, representation and visualization of in vivo diffusion processes using MRI.
He is coordinating the data acquisition and refinement sub-project in Seeing organ function - patient specific image data driven organ exploration and simulation, a large strategic project supported by The Knut and Alice Wallenberg Foundation.
