Abstract
We propose a technique for tissue characterization using instant features of intravascular ultrasound (IVUS) radiofrequency signals. In the proposed technique, the modular network self-organizing map (mnSOM) is used to classify the instant patterns of RF signals at short time. The conventional approaches have achieved the tissue characterization by the static information of the frequency spectrums. By contrast, in the proposed method, a time history of the frequency spectrum is used. In this paper, we describe the proposed method in brief and the results of comparative experiments are given.
