Segmentation of Calcified Plaques in Intravascular Ultrasound Images

Abstract

Intravascular ultrasound (IVUS) imaging is mostly used in the diagnosis and treatment of coronary artery diseases, especially in atherosclerosis, because it becomes very difficult to identify in the calcified regions manually. The IVUS images allow to visualize the inner portion of the coronary artery with enhanced resolution and also to acquire the cross-sectional images of arteries. Therefore, this paper presents a computational framework to identify the calcified region in IVUS images. In this paper, spatial fuzzy C-means approach is used to extract the exact boundary of the calcified plaque region in the IVUS images along with the wavelet transform decomposition. This clustering approach is capable of incorporating additional spatial information obtained from the neighboring pixels and also overcoming the limitations of noise and artifacts in IVUS coronary images. Several experiments have been performed on the different IVUS data and their experimental results are analyzed in terms of both quantitative and qualitative manner. The results revealed that the spatial fuzzy C-means provides better segmentation accuracy by extracting the calcified region as compared with other approaches.