We propose an algorithm applied to an intravascular ultrasound (IVUS) -based tissue characterization of coronary plaque. In the proposed algorithm, the probability density of each plaque is calculated by Kernel density estimation. The probability density varies continuously in the feature space. The present algorithm enables to give more delicate information to a medical doctor. In this paper, the proposed algorithm is described in brief and its effectiveness is confirmed by applying it to the read IVUS data.
