Automatic Framework for the Detection of Coronary Artery Calcification in IVUS Images
Abstract
The paper propose an automatic framework for the detection of coronary artery calcification in intravascular ultrasound (IVUS) images using texture analysis method. The texture features used is called Histogram of Equivalent Patterns (HEPs) Features. Experiments was conducted using 2175 IVUS images, 530 with calcification plague and 1645 without calci-fication plague. The images are from dataset B of MICCAI challenge 2011. The classifier used is 1-NN classifier. A 2-fold cross-validation process is applied to the IVUS image database to evaluate the performance of the proposed framework. The highest accuracy obtained is 95.89 %, using a variant of Com-pleted Local Binary Patterns (CLBP) descriptors as the features.