根據世界衛生組織2018年的數據統計，肝臟疾病已是全球器官組織疾病中第四大的死亡原因。在2015年時，因病毒性肝炎而死亡的人數高達134萬人，其中有大約72萬例死亡與慢性肝炎有關，病毒性肝炎同時也是誘發肝細胞癌和肝硬化的主要原因。
臨床上，肝臟疾病的診斷通常藉由肝臟病理切片進行，因為病理組織切片能提供細胞層級的詳細資訊，而肝炎的分級和分期通常則使用Ishak評分系統進行評估。但是，在對肝臟病理組織切片進行分期時，病理學家必須在顯微鏡或數位圖像下以不同放大倍率檢查整個肝臟組織切片的特徵。整個診斷過程非常耗時與費力，並且有因人為專業知識錯誤而導致診斷偏差的可能性存在。因此，病毒性肝炎的自動化診斷是極為重要的議題。
在Ishak評分系統中，纖維化分期是基於以下標準做為分期準則：門脈區域是否有纖維擴張；門脈是否與其他門脈區域或與中心靜脈區域橋接；或是否組織切片出現結節甚至肝硬化的特徵。根據上述纖維化分期標準，進行自動化分期之前，都必須先行判別出門靜脈區域和中央靜脈區域。一般而言，門脈區域和中心靜脈區域具有不同的結構特徵：即膽管僅會出現在門脈區域中，因此藉由膽管的偵測有助於判別門脈區域與中央靜脈區域。膽管的特徵型態多變，通常難以在單一倍率下檢測，因此病理學家會以不同倍率來觀察切片中膽管本身內部結構資訊或是其周圍資訊用以準確判斷。以此為前提，本研究提出了一個多尺度輸入的注意力機制卷積網絡，以模擬人類在肝臟病理切片中不同倍率下對膽管的檢測，並與其他現有的網路架構相比更能成功地分割出膽管。

Liver disease was the fourth cause of death in organ diseases worldwide, evaluated by the World Health Organization in 2018. Besides, according to the Global Hepatitis Report presented, the total number of deaths due to viral hepatitis in 2015 was approximately 1.34 million, of which about 720,000 deaths were related to chronic hepatitis. Viral hepatitis is also the main cause of hepatocellular carcinoma and cirrhosis.
Clinically, the diagnosis of liver disease is generally performed using liver biopsy since it provides a detailed information at the cellular level. The grading and staging of hepatitis are usually evaluated by the Ishak Score system. However, when staging a liver biopsy, the pathologists have to examine the characteristics of the entire liver biopsy at different magnifications under a microscope or digital images. The entire diagnosis process is quite time-consuming and laborious, and diagnostic deviation may occur as a result of expertise with potential human error. Therefore, the automatic diagnose of viral hepatitis is an extremely important issue.
In Ishak Score system, the fibrosis staging is assigned based on the findings that whether portal areas have fibrous expansion, bridging with other portal areas, bridging with the central veins, the appearance of nodules, and even cirrhosis. According to the above fibrosis staging criterion, before automatic staging, the portal areas and the central veins must be distinguished first. Generally, the portal areas and the central veins have different structural characteristics that bile ducts only appear in the portal areas. Hence the detection of the bile ducts can help distinguish the portal area from the central vein. The characteristics of the bile ducts have variant types that are usually difficult to detect under a single magnification. Therefore, the pathologists observe the internal structure information of the bile duct or the surrounding information in different magnifications for accurate detection. In this study, a multi-scale convolutional network with attention mechanism is proposed to simulate human examination of the bile ducts under different magnifications in liver biopsy, and successfully segment the bile ducts than other existing networks.

[1] World Health Organization. Global hepatitis report 2017. World Health Organization, 2017.
[2] Asrani, Sumeet K., et al. "Burden of liver diseases in the world." Journal of hepatology 70.1 (2019): 151-171.
[3] Villanueva, Augusto. “Hepatocellular Carcinoma.” The New England journal of medicine vol. 380,15 (2019): 1450-1462.
[4] Ishak, Kamal. "Histological grading and staging of chronic hepatitis." J hepatol 22 (1995): 696-699.
[5] Krizhevsky, Alex, Ilya Sutskever, and Geoffrey E. Hinton. "Imagenet classification with deep convolutional neural networks." Advances in neural information processing systems. 2012.
[6] Simonyan, Karen, and Andrew Zisserman. "Very deep convolutional networks for large-scale image recognition." arXiv preprint arXiv:1409.1556 (2014).
[7] Long, Jonathan, Evan Shelhamer, and Trevor Darrell. "Fully convolutional networks for semantic segmentation." Proceedings of the IEEE conference on computer vision and pattern recognition. 2015.
[8] Girshick, Ross, et al. "Rich feature hierarchies for accurate object detection and semantic segmentation." Proceedings of the IEEE conference on computer vision and pattern recognition. 2014.
[9] Goodfellow, Ian, et al. "Generative adversarial nets." Advances in neural information processing systems. 2014.
[10] Toshev, Alexander, and Christian Szegedy. "Deeppose: Human pose estimation via deep neural networks." Proceedings of the IEEE conference on computer vision and pattern recognition. 2014.
[11] Ronneberger, Olaf, Philipp Fischer, and Thomas Brox. "U-net: Convolutional networks for biomedical image segmentation." International Conference on Medical image computing and computer-assisted intervention. Springer, Cham, 2015.
[12] Badrinarayanan, Vijay, Alex Kendall, and Roberto Cipolla. "Segnet: A deep convolutional encoder-decoder architecture for image segmentation." IEEE transactions on pattern analysis and machine intelligence 39.12 (2017): 2481-2495.
[13] Chen, Liang-Chieh, et al. "Semantic image segmentation with deep convolutional nets and fully connected crfs." arXiv preprint arXiv:1412.7062 (2014).
[14] Chen, Liang-Chieh, et al. "Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs." IEEE transactions on pattern analysis and machine intelligence 40.4 (2017): 834-848.
[15] Chen, Liang-Chieh, et al. "Rethinking atrous convolution for semantic image segmentation." arXiv preprint arXiv:1706.05587 (2017).
[16] Chen, Liang-Chieh, et al. "Encoder-decoder with atrous separable convolution for semantic image segmentation." Proceedings of the European conference on computer vision (ECCV). 2018.
[17] Haris, Kostas, et al. "Hybrid image segmentation using watersheds and fast region merging." IEEE Transactions on image processing 7.12 (1998): 1684-1699.
[18] Wu, Hai-Shan, et al. "Segmentation of microscopic images of small intestinal glands with directional 2-d filters." Analytical and quantitative cytology and histology 27.5 (2005): 291.
[19] Matula, Petr, et al. "Single‐cell‐based image analysis of high‐throughput cell array screens for quantification of viral infection." Cytometry Part A: The Journal of the International Society for Advancement of Cytometry 75.4 (2009): 309-318.
[20] Simsek, Ahmet Cagri, et al. "Multilevel segmentation of histopathological images using cooccurrence of tissue objects." IEEE transactions on biomedical engineering 59.6 (2012): 1681-1690.
[21] Veta, Mitko, et al. "Automatic nuclei segmentation in H&E stained breast cancer histopathology images." PloS one 8.7 (2013): e70221.
[22] Irshad, Humayun, et al. "Methods for nuclei detection, segmentation, and classification in digital histopathology: a review—current status and future potential." IEEE reviews in biomedical engineering 7 (2013): 97-114.
[23] Kong, Jun, et al. "Machine-based morphologic analysis of glioblastoma using whole-slide pathology images uncovers clinically relevant molecular correlates." PloS one 8.11 (2013): e81049.
[24] Webster, J. D., and R. W. Dunstan. "Whole-slide imaging and automated image analysis: considerations and opportunities in the practice of pathology." Veterinary pathology 51.1 (2014): 211-223.
[25] Xing, Fuyong, and Lin Yang. "Robust nucleus/cell detection and segmentation in digital pathology and microscopy images: a comprehensive review." IEEE reviews in biomedical engineering 9 (2016): 234-263.
[26] Wang, Haibo, et al. "Mitosis detection in breast cancer pathology images by combining handcrafted and convolutional neural network features." Journal of Medical Imaging 1.3 (2014): 034003.
[27] Cui, Lei, et al. "A deep learning-based framework for lung cancer survival analysis with biomarker interpretation." BMC bioinformatics 21.1 (2020): 1-14.
[28] Kurc, Tahsin, et al. "Segmentation and Classification in Digital Pathology for Glioma Research: Challenges and Deep Learning Approaches." Frontiers in Neuroscience 14 (2020).
[29] Liu, Yun, et al. "Detecting cancer metastases on gigapixel pathology images." arXiv preprint arXiv:1703.02442 (2017).
[30] Song, Youyi, et al. "Accurate segmentation of cervical cytoplasm and nuclei based on multiscale convolutional network and graph partitioning." IEEE Transactions on Biomedical Engineering 62.10 (2015): 2421-2433.
[31] Huang, Wei-Che, et al. "Automatic HCC detection using convolutional network with multi-magnification input images." 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS). IEEE, 2019.
[32] Sayıcı, Mehmet Burak, Rikiya Yamashita, and Jeanne Shen. "Analysis Of Multi Field Of View Cnn And Attention Cnn On H&E Stained Whole-slide Images On Hepatocellular Carcinoma." arXiv preprint arXiv:2002.04836 (2020).
[33] He, Kaiming, et al. "Deep residual learning for image recognition." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.
[34] Yang, Maoke, et al. "Denseaspp for semantic segmentation in street scenes." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2018.
[35] Lin, Tsung-Yi, et al. "Focal loss for dense object detection." Proceedings of the IEEE international conference on computer vision. 2017.
[36] Nair, Vinod, and Geoffrey E. Hinton. "Rectified linear units improve restricted boltzmann machines." ICML. 2010.
[37] Ioffe, Sergey, and Christian Szegedy. "Batch normalization: Accelerating deep network training by reducing internal covariate shift." arXiv preprint arXiv:1502.03167 (2015).