Department of Electronic and Communication Engineering, North China Electric Power University, Baoding, P.R.China


Introduction: Intravascular ultrasound (IVUS) imaging is a widely accepted interventional imaging modality in clinical diagnosis and treatment of vascular diseases, especially coronary artery diseases. Accurate identification and classification of atherosclerotic plaques are essential for predict the possible evolution of plaques.

Materials and methods: Automated tissue characterization including atherosclerotic plaques, vessel branching and stent struts in IVUS gray-scale images is addressed. The texture features are firstly detected with local binary pattern (LBP), Haar-like and Gabor filter. Then, a Gentle Adaboost classifier is designed to classify tissue texture features. The manual characterization results obtained by experienced physicians were adopted as the golden standard to evaluate the accuracy.

Results: Results with clinically acquired image data indicate that the recognition accuracy of lipidic plaques can reach 94.54%, while classification precision of fibrous and calcified plaques can reach 93.08%. The recognition accuracy of branching and stents can reach 93.20% and 93.50%, respectively.

Conclusion: Fully automated identification of soft/hard plaques, vessel branching and stent struts can be realized through detecting and classifying the texture features of IVUS images. However, accurate classification of lipidic, fibrous, calcified and mixed plaques may be implemented according to grey-scale images combined with raw radio-frequency signals.


Intravascular ultrasound (IVUS), tissue characterization, texture features, Gentle Adaboost.