Mitsubishi Electric Research Laboratories, Inc.

The University of Arizona

Siemens Corporation

Rajagopalan Sundaresan

Mariappan S Nadar

Fatih M Porikli

Ali Bilgin

Yookyung Kim

A method for reconstructing a digital image from a set of measurements includes providing a previous image frame in a time series of measurements of an image signal and a current image frame in the time series, calculating an estimated motion vector for a spatial point and current time point between the previous and current image frames, calculating a motion compensated current image frame from the previous image frame, estimating a known support set of a sparse signal estimate of the motion compensated current image frame where the support set comprises indices of non-zero elements of the sparse signal estimate, calculating a sparse signal corresponding to the current image frame whose support contains a smallest number of new additions to the known support set while satisfying a data consistency constraint, and correcting the motion compensated current image frame image frame from the sparse signal.

System and method for motion-compensated compressed sensing for dynamic imaging

A method reduces multiplicative and additive noise in image pixels by clustering similar patches of the pixels into clusters. The clusters form nodes in an affinity net of nodes and vertices. From each cluster, a dictionary is learned by a sparse combination of corresponding atoms in the dictionaries. The patches are aggregated collaboratively using the dictionaries to construct a denoised image.

Image filtering by sparse reconstruction on affinity net

Growing community of inventors

Tucson, AZ, United States of America

Rajagopalan Sundaresan