肝癌的病理組織分化程度分級在癌症預後及術後追蹤上是一項重要的步驟。然而，傳統的診斷方式主觀且耗時。因此本論文提出一種自動化將肝臟切片影像分為四種分級的其中一種的方法。此方法使用多倍率影像卷積神經網路作為特徵擷取器來取得難以量化之紋理特徵。並使用基於影像處理的技巧取得關於血管、肝細胞、肝小梁的特徵。最後，建構一個分類器使用以上所有特徵預測肝癌影像的癌症分級。分類器在病理影像肝癌分級中達到了 98.2%的預測準確度。根據實驗結果顯示，分類器對肝癌病理影像分級預測展現了可靠的準確度，並證明來自影像卷積神經網路的特徵及來自肝小梁的特徵皆能輔助分類器的判斷，提升分級準確度。

The accurate grading of Hepatocellular Carcinoma (HCC) in histopathological liver tissue images is crucial to prognosis and treatment planning. However, the traditional diagnosis process is subjective and time-consuming. Accordingly, this study proposes a novel method to automatically classify a liver tissue as one of four different tumor grades, namely 1(Well differentiated), 2(Moderately differentiated),3(Poorly differentiated), and (Undifferentiated). In the proposed method, a CNN-based feature extractor is trained to retrieve the features of the tissue that are hard to be quantized by traditional method. In addition, sinusoid, cell and trabecular features are extracted by image processing. Finally, a classifier is constructed using all the features to predict the tumor grade of the input tissue.
The proposed classifier reaches a 98.2% accuracy on whole slide images (WSIs) of HCC.
The experimental results show that the proposed method performs well on liver tumor grading prediction and proves that the features obtained from the CNN-based feature extractor and trabecular features can further enhance the performance of the classifier.

[1] Ministry of Health and Welfare of Taiwan, the Death Statistics of Taiwanese in
2017. Retrieved March 8, 2019, from https://www.mohw.gov.tw/cp-16-41794
1.html
[2] Cancer Treatment Centers of America, Types of Liver Cancer: Common, Rare,
and More Varieties. Retrieved March 8, 2019, from
https://www.cancercenter.com/cancer-types/liver-cancer/types
[3] S. Petushi, F. U. Garcia, M. M. Haber, C. Katsinis, and A. Tozeren, “Large-scale
computations on histology images reveal grade-differentiating parameters for
breast cancer,” BMC Med. Imaging, vol. 6, no. 1, p. 14, 2006.
[4] S. Naik, S. Doyle, S. Agner, A. Madabhushi, M. Feldman, and J. Tomaszewski,
“Automated gland and nuclei segmentation for grading of prostate and breast
cancer histopathology,” in 2008 5th IEEE International Symposium on Biomedical
Imaging: From Nano to Macro, Proceedings, ISBI, 2008, pp. 284–287.
[5] S. Doyle, S. Agner, A. Madabhushi, M. Feldman, and J. Tomaszewski,
“Automated grading of breast cancer histopathology using spectral clustering with
textural and architectural image features,” in 2008 5th IEEE International
Symposium on Biomedical Imaging: From Nano to Macro, Proceedings, ISBI,
2008, pp. 496–499.
[6] Lai, Y.H. & W. Huang, P & L. Lin, P. (2007). An effective automated grading system for HCC in biopsy images. 396-402.
[7] S. I. Niwas, P. Palanisamy, W. J. Zhang, N. A. Mat Isa, and R. Chibbar, “Log
gabor wavelets based breast carcinoma classification using least square support
vector machine,” in 2011 IEEE International Conference on Imaging Systems and Techniques, IST 2011 - Proceedings, 2011, pp. 219–223.
[8] C. Atupelage, H. Nagahashi, M. Yamaguchi, T. Abe, A. Hashiguchi and M.
Sakamoto, "Multifractal feature descriptor for grading Hepatocellular
carcinoma," Proceedings of the 21st International Conference on Pattern
Recognition (ICPR2012), Tsukuba, 2012, pp. 129-132.
[9] Atupelage C, Nagahashi H, Kimura F, et al. Computational hepatocellular carcinoma tumor grading based on cell nuclei classification. Journal of Medical Imaging. 2014;1(3):034501. doi:10.1117/1.JMI.1.3.034501.
[10] Aziz MA, Kanazawa H, Murakami Y, Kimura F, Yamaguchi M, Kiyuna T, et al. Enhancing automatic classification of hepatocellular carcinoma images through image masking, tissue changes and trabecular features. J Pathol Inform 2015;6:26.
[11] Kiyuna, T., Saito, A., Marugame, A., Yamashita, Y., Ogura, M., Cosatto, E., ... Sakamoto, M. (2013). Automatic classification of hepatocellular carcinoma images based on nuclear and structural features. In Progress in Biomedical Optics and Imaging - Proceedings of SPIE (Vol. 8676). [86760Y] DOI: 10.1117/12.2006667
[12] Bingbing Xia, Huiyan Jiang, Huiling Liu, and Dehui Yi, “A Novel Hepatocellular Carcinoma Image Classification Method Based on Voting Ranking Random Forests,” Computational and Mathematical Methods in Medicine, vol. 2016, Article ID 2628463, 8 pages, 2016
[13] P. W. Huang and Y. H. Lai, “Effective segmentation and classification for HCC
biopsy images,” Pattern Recognit., vol. 43, no. 4, pp. 1550–1563, 2010.
[14] Ishikawa M, Ahi ST, Kimura F, Yamaguchi M, Nagahashi H, Hashiguchi A, et
al. Segmentation of sinusoids in hematoxylin and eosin stained liver specimens
using an orientation-selective filter. Open J Med Imaging. 2013;3:144–55.
[15] Ishikawa M, Murakami Y, Ahi ST, et al. Automatic quantification of
morphological features for hepatic trabeculae analysis in stained liver specimens.
J Med Imaging (Bellingham). 2016;3(2):027502.
[16] Ishikawa M, Murakami Y, Ahi ST, Yamaguchi M, Kobayashi N, Kiyuna T,
Yamashita Y, Saito A, Abe T, Hashiguchi A, Sakamoto M (2016) Automatic
quantification of morphological features for hepatic trabeculae analysis in
stained liver specimens. J Med Imag. doi: 10.1117/1.JMI.3.2.027502
[17] J. Long, E. Shelhamer, and T. Darrell. Fully convolutional networks for semantic
segmentation. InCVPR,2015
[18] V. Badrinarayanan, A. Kendall, and R. Cipolla. Segnet: Adeep convolutional
encoder-decoder architecture for image segmentation n.arXiv:1511.00561v2
[cs.CV], 2015.
[19] K. Simonyan and A. Zisserman. Very deep convolutional networks for large
scale image recognition. In ICLR, 2015
[20] Y. Zhou, H. Chang, K. Barner, P. Spellman, and B. Parvin, “Classification of
histology sections via multispectral convolutional sparse coding,” in Proceedings
of the IEEE Computer Society Conference on Computer Vision and Pattern
Recognition, 2014, pp. 3081–3088.
[21] Y. Xu, Z. Jia, Y. Ai, F. Zhang, M. Lai, and E. I. C. Chang, “Deep convolutional
activation features for large scale Brain Tumor histopathology image classification
and segmentation,” in ICASSP, IEEE International Conference on Acoustics,
Speech and Signal Processing - Proceedings, 2015, pp. 947–951.
[22] Krizhevsky, I. Sutskever, and G. Hinton. Imagenet classificationwith deep
convolutional neural networks. In NIPS, 2012
[23] A. Cruz-Roa et al., “Accurate and reproducible invasive breast cancer detection in
whole-slide images: A Deep Learning approach for quantifying tumor extent,” Sci.
Rep., vol. 7, no. 46450, 2017.
[24] D. Cires ̧an, A. Giusti, L. Gambardella, and J. Schmidhuber. Mi-tosis detection
in breast cancer histology images with deep neuralnetworks. InMICCAI, 2013
[25] Li S, Jiang H, Pang W (2017) Joint multiple fully connected convolutional neural
network with extreme learning machine for hepatocellular carcinoma nuclei
grading. Comput Biol Med 84:156–167
[26] Wei-Che Huang, Automatic HCC Detection Using Convolutional Network with
Multi-Magnification Input Images. AICAS 2019
[27] Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean.Efficient estimation
of word representationsin vector space.ICLR Workshop, 2013
[28] Chen, T. & Guestrin, C. Xgboost: A scalable tree boosting system. In Proc. 22nd
ACM SIGKDD International Conference on Knowledge Discovery and Data
Mining 785–794 (ACM, 2016)
[29] Liu, J.; Dang, H.; Wang, X.W. The significance of intertumor and intratumor
heterogeneity in liver cancer. Exp. Mol. Med. 2018, 50, e416