Multiscale motion mapping

In medical imaging, analysis of the heart and its motion is of considerable importance. Historically, this has been performed by "eye balling", i.e., by subjective assessment, by measurements on still frames and M-Modes images, and later by Doppler techniques that are based on the phase-shift of a reflected ultrasound signal and are by necessity constrained to the motion velocity component parallel to the ultrasound beam, while motion perpendicular to the beam will produce no phase shift in the reflected ultrasound beam and therefore appear as "zero velocity" by such Dppler techniques.

To overcome these problems, a solution was sought based on mathematical imaging by an academical consortium consisting of cardiovascular researchers at the University Hospital of Basel and mathematicians at the Ecole Polytechnique Federale of Lausanne, Switzerland. This resulted in a technique called Multiscale Motion Mapping because it relies on a combination multiscale image analysis, motion and deformation analysis using mathematical moments, and linear algebra, and uses 2D vectors with lengths proportional to velocity or deformation for visual velocity display.

Soon after the publication in international conventions and journal articles, similar techniques based on these ideas have been implemented by the medical imaging industry, e.g. by Siemens under the designation "Vector Velocity Imaging" (2006).