隨著現代飲食文化的進步，愈來愈多的人口具有高血壓、肥胖等問題，間接造成心血管疾病的形成。心血管疾病的形成主要是來自於動脈粥狀硬化，動脈粥狀硬化是一個緩慢斑塊堆積的過程，一方面經由血液的流動而造成斑塊破裂，另一方面減少腦部血液供給而造成腦中風及心肌梗塞等疾病。血管內超音波目前為臨床上用來診斷動脈粥狀硬化的方式，因其具有即時成像與良好穿透性，在影像上提供較多血管壁的結構與組成資訊。現今商用影像系統提供一項藉由超音波回波訊號來提取訊號中的頻域資訊進行斑塊辨識的技術，一方面提供臨床上一個參考依據，但另一方面臨床醫師認為在斑塊辨識上仍有可以改善的空間。因此，此研究提出一個人工智慧技術應用在分割灰階血管內超音波影像的中膜-外膜邊界、管腔區域，並且定位鈣化斑塊堆積區域。
模型的影像訓練資料是由彰化秀傳紀念醫院所提供，來自於18位受測者，總計為713張影像。其中每一張灰階血管內超音波影像資料由人工標記中膜-外膜所包圍面積、管腔區域及鈣化堆積的位置。在方法上，主要是利用深度學習中的卷積神經網路，並且配合疊加網路的概念，將三個神經網路互相連接，降低網路錯誤辨識非斑塊堆積區域的組織。此外，訓練過程中利用三種不同數學意義的損失函數修改模型內部的神經元之間的權重。在準確度的計算上使用Dice score、精確率(Precision)、召回率(Recall)及特異性(Specificity)來評估性能。
在模型訓練過程中，由於斑塊堆積的面積及鈣化組織的位置與斑塊的穩定性有關，實驗主要是關注在斑塊區域與鈣化的辨識正確性。從實驗結果發現，此研究所提供的模型分割演算法在不同損失函數下，都能夠辨識斑塊區域達到高於90%的準確度，雖然鈣化的準確度落在67%，但相較於商用軟體與人工比較能夠提供高於兩倍的正確性。因此，此研究結果可以提供更加完善的辨識資訊，以利於後續的治療。

With the advance of food culture in modern societies, there are more and more people suffer from hypertension and diabetes may result in cardiovascular diseases. The main factor for cardiovascular diseases is atherosclerosis which is a slow process and would finally lead to severe symptoms like myocardial infarction or brain stroke due plaque rupture or reduction of blood supply to the heart system or brain. Intravascular ultrasound (IVUS) imaging has been a common technique to diagnose atherosclerosis in clinical application. Owing to its characteristics of real-time and well penetration into vessel walls, IVUS imaging provides more geometrical information and composition of blood vessel walls. Inside recent commercial IVUS imaging system, the software named as Virtual Histology IVUS (VH-IVUS) imaging utilizes the backscattering radiofrequency ultrasound signals to attain its featured information from frequency domain. This technique on one hand is proved to correctly characterize plaque types into calcium, fibrous, fibro-fatty and necrotic cores and provides a referenced information in clinics. However, on the other hand, cardiologists consider that there is still improvement space on plaque characterization technique. Therefore, this study demonstrates a method based on artificial intelligence to segment borders of media-adventitia and luminal region as well as locations of calcified tissues.
The in-vivo image dataset is provided from Show Chwan Memorial Hospital, Changhua, Taiwan and the number of images is 713 from 18 subjects with atherosclerosis. In the dataset, each grayscale IVUS image is labeled with classes of MA border, lumen and calcium. The proposed method utilized the method of convolutional neural network in deep learning technique incorporated with the concept of cascaded network to reduce the occurrence of incorrectly detection on the regions outside plaque burden by connecting triple neural networks. Besides, in the course of learning process, three loss functions with different mathematical properties are used to adapt the weighting parameters between neurons in convolutional networks. The evaluation measurement is implemented with Dice score, precision, recall and specificity to estimate the performance of the proposed method.
In the duration of learning procedure, the region of interest is focused on accurate detection of the plaque regional size and calcified tissues since the stabilization or vulnerability of plaque burden is relevant to plaque region and location of calcium. From the experiments, the method proposed in this study could reach the high accuracy over 0.9 with the usage of various loss functions. Although the accuracy in detection of calcium is located at about 0.67, it could provide twice better accurate information in contrast to the results from VH-IVUS.
These results could provide better characterized information for latter treatment strategies of atherosclerosis.

[1] J. Sanz and Z. A. Fayad, "Imaging of atherosclerotic cardiovascular disease," Nature, vol. 451, no. 7181, pp. 953-957, 2008.
[2] R. Fedewa et al., "Artificial Intelligence in Intracoronary Imaging," Current Cardiology Reports, vol. 22, p. 46, 2020.
[3] S. E. Nissen and P. Yock, "Intravascular ultrasound: novel pathophysiological insights and current clinical applications," Circulation, vol. 103, no. 4, pp. 604-616, 2001.
[4] A. Katouzian, E. D. Angelini, S. G. Carlier, J. S. Suri, N. Navab, and A. F. Laine, "A state-of-the-art review on segmentation algorithms in intravascular ultrasound (IVUS) images," IEEE Transactions on Information Technology in Biomedicine, vol. 16, no. 5, pp. 823-834, 2012.
[5] L. S. Athanasiou, D. I. Fotiadis, and L. K. Michalis, Atherosclerotic Plaque Characterization Methods Based on Coronary Imaging. Academic Press, 2017.
[6] G. S. Mintz, "Intravascular imaging of coronary calcification and its clinical implications," JACC: Cardiovascular Imaging, vol. 8, no. 4, pp. 461-471, 2015.
[7] A. Nair, B. D. Kuban, E. M. Tuzcu, P. Schoenhagen, S. E. Nissen, and D. G. Vince, "Coronary plaque classification with intravascular ultrasound radiofrequency data analysis," Circulation, vol. 106, no. 17, pp. 2200-2206, 2002.
[8] C. Von Birgelen et al., "Morphometric analysis in three-dimensional intracoronary ultrasound: an in vitro and in vivo study performed with a novel system for the contour detection of lumen and plaque," American heart journal, vol. 132, no. 3, pp. 516-527, 1996.
[9] M. E. Plissiti, D. I. Fotiadis, L. K. Michalis, and G. E. Bozios, "An automated method for lumen and media-adventitia border detection in a sequence of IVUS frames," IEEE Transactions on Information Technology in Biomedicine, vol. 8, no. 2, pp. 131-141, 2004.
[10] M. Kass, A. Witkin, and D. Terzopoulos, "Snakes: Active contour models," International journal of computer vision, vol. 1, no. 4, pp. 321-331, 1988.
[11] Y. Zhu and Z. Yan, "Computerized tumor boundary detection using a Hopfield neural network," IEEE transactions on medical imaging, vol. 16, no. 1, pp. 55-67, 1997.
[12] D. Gil, P. Radeva, and J. Saludes, "Segmentation of artery wall in coronary IVUS images: A probabilistic approach," in Proceedings 15th International Conference on Pattern Recognition. ICPR-2000, 2000, vol. 4, pp. 352-355: IEEE.
[13] D. Gil, P. Radeva, J. Saludes, and J. Mauri, "Automatic segmentation of artery wall in coronary IVUS images: a probabilistic approach," in Computers in Cardiology 2000. Vol. 27 (Cat. 00CH37163), 2000, pp. 687-690: IEEE.
[14] A. Taki, A. Roodaki, S. K. Setarehdan, S. Avansari, G. Unal, and N. Navab, "An IVUS image-based approach for improvement of coronary plaque characterization," Computers in biology and medicine, vol. 43, no. 4, pp. 268-280, 2013.
[15] X. Tang, "Texture information in run-length matrices," IEEE transactions on image processing, vol. 7, no. 11, pp. 1602-1609, 1998.
[16] C.-c. Chang and C.-j. Lin, "Libsvm: A library for support vector machines. software available at http://www. csie. ntu. edu. tw/cjlin/libsvm," 2001.
[17] S. Escalera, O. Pujol, J. Mauri, and P. Radeva, "Intravascular ultrasound tissue characterization with sub-class error-correcting output codes," Journal of Signal Processing Systems, vol. 55, no. 1-3, pp. 35-47, 2009.
[18] S. Su, Z. Hu, Q. Lin, W. K. Hau, Z. Gao, and H. Zhang, "An artificial neural network method for lumen and media-adventitia border detection in IVUS," Computerized Medical Imaging and Graphics, vol. 57, pp. 29-39, 2017.
[19] J. Yang, L. Tong, M. Faraji, and A. Basu, "IVUS-Net: an intravascular ultrasound segmentation network," in International Conference on Smart Multimedia, 2018, pp. 367-377: Springer.
[20] J. Long, E. Shelhamer, and T. Darrell, "Fully convolutional networks for semantic segmentation," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 3431-3440.
[21] C. Szegedy et al., "Going deeper with convolutions," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 1-9.
[22] T. C. Ulli and D. Gupta, "Segmentation of Calcified Plaques in Intravascular Ultrasound Images," in Smart Computing Paradigms: New Progresses and Challenges: Springer, 2020, pp. 57-67.
[23] K.-S. Chuang, H.-L. Tzeng, S. Chen, J. Wu, and T.-J. Chen, "Fuzzy c-means clustering with spatial information for image segmentation," computerized medical imaging and graphics, vol. 30, no. 1, pp. 9-15, 2006.
[24] A. Krizhevsky, I. Sutskever, and G. E. Hinton, "Imagenet classification with deep convolutional neural networks," in Advances in neural information processing systems, 2012, pp. 1097-1105.
[25] K. Simonyan and A. Zisserman, "Very deep convolutional networks for large-scale image recognition," arXiv preprint arXiv:1409.1556, 2014.
[26] H. Noh, S. Hong, and B. Han, "Learning deconvolution network for semantic segmentation," in Proceedings of the IEEE international conference on computer vision, 2015, pp. 1520-1528.
[27] O. Ronneberger, P. Fischer, and T. Brox, "U-net: Convolutional networks for biomedical image segmentation," in International Conference on Medical image computing and computer-assisted intervention, 2015, pp. 234-241: Springer.
[28] V. Badrinarayanan, A. Kendall, and R. Cipolla, "Segnet: A deep convolutional encoder-decoder architecture for image segmentation," IEEE transactions on pattern analysis and machine intelligence, vol. 39, no. 12, pp. 2481-2495, 2017.
[29] T.-Y. Lin, P. Dollár, R. Girshick, K. He, B. Hariharan, and S. Belongie, "Feature pyramid networks for object detection," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 2117-2125.
[30] P. F. Christ et al., "Automatic liver and tumor segmentation of CT and MRI volumes using cascaded fully convolutional neural networks," arXiv preprint arXiv:1702.05970, 2017.
[31] K. Shung, "Diagnostic ultrasound: Imaging and Blood flow," United State of America, 2006.
[32] I. Goodfellow, Y. Bengio, and A. Courville, Deep learning. MIT press, 2016.
[33] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, "Gradient-based learning applied to document recognition," Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, 1998.
[34] O. Russakovsky et al., "Imagenet large scale visual recognition challenge," International journal of computer vision, vol. 115, no. 3, pp. 211-252, 2015.
[35] S. Ioffe and C. Szegedy, "Batch normalization: Accelerating deep network training by reducing internal covariate shift," arXiv preprint arXiv:1502.03167, 2015.
[36] K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770-778.
[37] N. Ketkar and E. Santana, Deep Learning with Python. Springer, 2017.
[38] T. A. Sorensen, "A method of establishing groups of equal amplitude in plant sociology based on similarity of species content and its application to analyses of the vegetation on Danish commons," Biol. Skar., vol. 5, pp. 1-34, 1948.
[39] L. R. Dice, "Measures of the amount of ecologic association between species," Ecology, vol. 26, no. 3, pp. 297-302, 1945.
[40] S. S. M. Salehi, D. Erdogmus, and A. Gholipour, "Tversky loss function for image segmentation using 3D fully convolutional deep networks," in International Workshop on Machine Learning in Medical Imaging, 2017, pp. 379-387: Springer.
[41] T.-Y. Lin, P. Goyal, R. Girshick, K. He, and P. Dollár, "Focal loss for dense object detection," in Proceedings of the IEEE international conference on computer vision, 2017, pp. 2980-2988.
[42] M. Abadi et al., "Tensorflow: A system for large-scale machine learning," in 12th {USENIX} symposium on operating systems design and implementation ({OSDI} 16), 2016, pp. 265-283.
[43] D. P. Kingma and J. Ba, "Adam: A method for stochastic optimization," arXiv preprint arXiv:1412.6980, 2014.
[44] Ö. Çiçek, A. Abdulkadir, S. S. Lienkamp, T. Brox, and O. Ronneberger, "3D U-Net: learning dense volumetric segmentation from sparse annotation," in International conference on medical image computing and computer-assisted intervention, 2016, pp. 424-432: Springer.
[45] C. H. Sudre, W. Li, T. Vercauteren, S. Ourselin, and M. J. Cardoso, "Generalised dice overlap as a deep learning loss function for highly unbalanced segmentations," in Deep learning in medical image analysis and multimodal learning for clinical decision support: Springer, 2017, pp. 240-248.
[46] F. Milletari, N. Navab, and S.-A. Ahmadi, "V-net: Fully convolutional neural networks for volumetric medical image segmentation," in 2016 fourth international conference on 3D vision (3DV), 2016, pp. 565-571: IEEE.
[47] H. Kervadec, J. Bouchtiba, C. Desrosiers, E. Granger, J. Dolz, and I. B. Ayed, "Boundary loss for highly unbalanced segmentation," in International conference on medical imaging with deep learning, 2019, pp. 285-296.
[48] C. Szegedy, S. Ioffe, V. Vanhoucke, and A. A. Alemi, "Inception-v4, inception-resnet and the impact of residual connections on learning," in Thirty-first AAAI conference on artificial intelligence, 2017.