Spatio-temporal summarizing method of periodic image sequences with Kohonen maps

Abstract

This work relates to the study of periodic events in medical imaging. Currently, biological phenomena exhibiting a periodic behaviour such as the heart motion are observed through the continuous recording of signals / images. Indeed, due to various reasons, cycle duration may slightly vary in duration and magnitude. It is important for understanding to be able to extract the meaningful information from the mass of acquired data. This paper presents a new neural-based method for the extraction of summarized cycle from long and massive recordings. Its concept is simple and it can be implemented on a hardware architecture to make the process very fast. The proposed method is demonstrated on sequences of noise-free and noisy synthetic images eventually in the presence of artefacts. © Springer-Verlag Berlin Heidelberg 2008.