Plaque Mass Border Detection and Classification in Intravascular Ultrasound Images Using Plaque Mass Weight-K-Nearest Neighbour Classifier

摘要

The problem of plaque mass detection as well as border detection in Intravascular Ultrasound Images (IVUS) has been well studied. Numerous algorithms have been named for the classification of plaque but suffer to achieve higher performance. To improve the performance in the detection of a plaque, an efficient Plaque Mass measure based border detection and classification algorithm has been studied. In general the presence of plaque has been identified by monitoring the vessel images. Towards that, the proposed method reads the image and preprocesses to remove noisy particles. Second stage, the image has been segmented according to gray scale values. Third, the feature of edges and plaque area has been identified to produce the feature vector. Finally, for each class the method estimates the plaque mass weight measure by applying PMW (Plaque Mass weight)-KNN (K-Nearest Neighbour) algorithm to produce a weight. Finally, the method has been measured for its efficiency and the algorithms have produced efficient results. The proposed algorithm improves the performance of plaque mass classification and reduces the false ratio.