Vulnerable plaque identification is important in coronary heart disease diagnosis. Intravascular Optical Coherence Tomography (IVOCT) is an imaging modality which can characterize the appearance of vulnerable plaques. However, current used manual reading of the images is time consuming and subjective. Therefore, an automated and objective assessment of the plaque is necessary. This paper proposes a
method for automatic identification of potential vulnerable plaque such as fibroatheroma in IVOCT images. In the proposed method, a graph search based method is applied to detect the region of interest (ROI) including the inner lumen border and outer border. Then various appearance features including the textures and the shape of ROI are extracted. A classifier is trained using support vector machine to detect the
presence of fibroath eroma plaque in the IVOCT images. The proposed method is evaluated using a dataset of 200 images from 24 different pullbacks. Experimental results show that the proposed method achieves a
mean accuracy of 90%, with sensitivity of 88% and specificity of 92%, in identifying fibroatheroma plaque in IVOCT images.