肝硬化是漸進性的疾病，主要歸因於肝臟長期損傷，其特徵在於纖維化組織會逐漸替代正常肝組織。長期且漸進的肝實質纖維化症狀可導致慢性肝病，並且可能會產生許多併發症，包括腹水，靜脈曲張，肝性腦病，肝細胞癌，肝肺綜合徵和凝血障礙。臨床上，肝硬化的診斷通常是基於血液檢查、切片檢查或是較為先進的醫療器具，例如彈性影像；然而目前電腦斷層仍然大規模的被應用在臨床上的影像檢查，特別是針對腹部的疾病。本研究分為兩部分: 一、本研究提出了一種新的電腦輔助診斷系統，會基於腹部電腦斷層掃描（CT）的四種數據自動診斷肝硬化，此四種數據紀錄注射顯影劑後在不同時間點的影像，包括未打顯影劑的電腦斷層影像，動脈相，延遲相和肝門靜脈相。此系統的開發是藉由自動肝臟分割方法和使用機器學習算法的分類方式，將肝硬化區分為輕度或嚴重水平。二、深度學習（DL）也應用於自動肝臟分割，以克服自定義算法上出現的困難，例如模糊的邊界，異質的外觀和高度變化的肝臟形狀，並且達到更準確的性能。方法:一、我們的數據分別來自義大醫院及成大醫院，首先，我們會計算CT數據的逆梯度圖，得出局部區域相對平滑的特徵，並且作為肝臟分割的理論基礎。然後通過比較每張切片的每個二元標記組的質心和面積，實現自動切割肝臟的目的。在分類步驟中，我們計算一階特徵和紋理相關特徵來描述肝實質的影像強度。同時，也量化VOI中的直方圖分佈和肝臟的輪廓形狀作為分類特徵。最後，包括分類樹，集合樹，訓練支持向量機（SVM）和神經網絡（NN）等分類器將病患的肝硬化做臨床分級。二、對於基於深度學習方法的肝臟分割，我們從名為”利用3D圖像重建比較算法數據庫（3D-IRCADb）”的網路資料庫，和來自成功大學放射腫瘤部的資料庫中獲取數據。該部分包含預處理和訓練步驟。在預處理步驟中，HU數據被轉換到可以強調肝臟特徵的影像域中。然後是訓練模型的步驟，我們用具有U-net結構的深度學習模型結合三種損失函數的來找出最合適的模型設置。我們使用Dice，結構相似性（SSIM）和峰值信噪比（PSNR）來評估性能。結果：一、在肝硬化的分類表現方面，基於集成學習算法對於脾臟相關特徵和紋理特徵進行的分類方式可以達到90％以上的準確性，靈敏度，特異性，證明了潛在的臨床適用性。二、我們使用具有五通道輸入格式的深度學習模型實現了最高性能。與手動分割相比，基於深度學習方法的自動肝臟分割的準確率可達到約88％。結論：雖然電腦斷層已經廣泛的被應用在急性的腹部疾病，目前的電腦輔助診斷系統主要還是基於血液檢查相關參數還有先進的醫療儀器設備，例如彈性影像及核磁共振影像。所以，我們開發了一項基於影像處理算法的自動化分割肝臟方法，並與機器學習演算法結合，以此來自動化偵測肝硬化；為了要解決傳統肝臟切割方式產生的問題，我們也應用了深度學習的方法。在這項研究中，我們證明了上述兩個題目有潛力去達到臨床可接受的準確率及臨床可用性。

Liver cirrhosis is a progressive disease which attributes to long-term damage and is characterized by the replacement from normal hepatic area to scar tissue. The symptoms of prolonged and increasing fibrosis of hepatic parenchyma can result in chronic liver disease and eventually come at a cost of many complications, such as varices, ascites, hepatic encephalopathy, hepatocellular carcinoma, hepatopulmonary syndrome, and coagulation disorders. In the clinic, the current computer-aided diagnosis (CAD) systems mainly depend on the blood test, biopsy or advanced imaging equipment such as elastography; however, computed tomography (CT) plays an important role in the clinical protocol, particularly for abdominal diseases. This study is divided into two parts. Firstly, a new computer-aided diagnosis system is proposed to automatically diagnose liver cirrhosis based on four-phase CT data, which record the Hounsfield Unit (HU) in four different periods after injection of contrast, including the non-contrast phase, the arterial phase, the delay phase, and the portal venous phase. We aimed to classify the cirrhosis into mild or severe levels by automatic liver segmentation method based on image processing methods and machine learning algorithms. Secondly, deep learning (DL) method is also developed for automatic liver segmentation in order to overcome the difficulties, such as ambiguous boundaries, heterogeneous appearances and highly varied shapes of the liver, which may occur in the previous segmentation method. Also, by means of the DL algorithm, we can achieve better and more accurate performance. Method: 1. In the beginning, we collected the data from both NCKU and Eda hospitals. Firstly, the gradient maps of CT data are inverted and used to derive the relatively smooth features in the local area, which served as the theoretical foundation of liver segmentation. Then the centroid and area of each binary labeled groups are compared from one slice to another to quantitatively extract the liver volume of interest (VOI) automatically. In the classification step, first-order features and texture-based features are calculated to describe the intensity arrangement of liver parenchyma. Some parameters are also used to quantify the distribution of intensity in VOI and the shape of miscellaneous contours of livers. Finally, the classifiers, including decision tree (DT), ensemble learning with tree templates, trained support vector machine (SVM) and neural network (NN) classifier, are used to classify the subjects into the malignant or benign stage. 2. As for the liver segmentation based on the deep learning method, we acquired the data from an online database, which is called the 3D Image Reconstruction for Comparison of Algorithm Database (3D-IRCADb) as well as the data from the NCKU Hospital. The protocol involves the preprocessing and training step. In the preprocessing step, Hounsfield unit windowing is applied to exclude the irrelevant parts of objects. Then follows the training step, where the U-net structure DL models with three kinds of loss functions are applied to determine the most suitable setting. We use Dice, Structure Similarity (SSIM) and Peak Signal-to-Noise Ratio (PSNR) to evaluate the performance. Results: 1. In terms of classification performance of liver cirrhosis, the application of the protocol based on ensemble learning algorithm with spleen-related features and texture features can achieve the accuracy, sensitivity, specificity all over 90%, demonstrating the potential clinical applicability of the proposed approach. 2. The DL model with 5-channel input data achieves the best performance. The accuracy of automatic liver segmentation based on the DL method can reach approximate 88% in comparison with manual segmentation. Conclusion: Althought CT has been widely applied in the diagnosis of acute abdominal diseases, the current CAD system mainly depends on the blood test, biopsy or advanced imaging equipment, such as elastography and MRI. In this study, We developed a CAD system for cirrhosis detection on CT images by using the automatic liver segmentation method based on traditional algorithms and machine learning method. We also establish a deep learning model to do automatic live segmentation in order to overcome the limitations when the traditional segmentation method is applied. In this study, we demonstrated that both of our projects have the potential to achieve clinically acceptable accuracy and clinical applicability.

[1] D. Y.Zhang and S. L.Friedman, “Fibrosis-dependent mechanisms of hepatocarcinogenesis,” Hepatology, vol. 56, no. 2. pp. 769–775, Aug-2012.
[2] S. K.Asrani, H.Devarbhavi, J.Eaton, and P. S.Kamath, “Burden of liver diseases in the world,” J. Hepatol., vol. 70, no. 1, pp. 151–171, Jan.2019.
[3] Y. A.Ghouri, I.Mian, andJ. H.Rowe, “Review of hepatocellular carcinoma: Epidemiology, etiology, and carcinogenesis.,” J. Carcinog., vol. 16, p. 1, 2017.
[4] A.Jemal, F.Bray, M. M.Center, J.Ferlay, E.Ward, and D.Forman, “Global cancer statistics,” CA. Cancer J. Clin., vol. 61, no. 2, pp. 69–90, Mar.2011.
[5] R. G.Knodell et al., “Formulation and application of a numerical scoring system for assessing histological activity in asymptomatic chronic active hepatitis,” Hepatology, vol. 1, no. 5, pp. 431–435, Sep.1981.
[6] ISHAK andK.G., “Chronic hepatitis: morphology and nomenclature,” Mod. Path., vol. 7, pp. 690–713, 1994.
[7] P.Bedossa and T.Poynard, “An algorithm for the grading of activity in chronic hepatitis C,” Hepatology, vol. 24, no. 2, pp. 289–293, Aug.1996.
[8] J.Westin, L. M.Lagging, R.Wejstål, G.Norkrans, and A. P.Dhillon, “Interobserver study of liver histopathology using the Ishak score in patients with chronic hepatitis C virus infection,” Liver Int., vol. 19, no. 3, pp. 183–187, Jun.1999.
[9] R.Rozario and B.Ramakrishna, “Histopathological study of chronic hepatitis B and C: a comparison of two scoring systems,” J. Hepatol., vol. 38, no. 2, pp. 223–229, Feb.2003.
[10] G.Sebastiani, “Non-invasive assessment of liver fibrosis in chronic liver diseases: implementation in clinical practice and decisional algorithms.,” World J. Gastroenterol., vol. 15, no. 18, pp. 2190–203, May2009.
[11] C.Jézéquel et al., “P0709 : Survival of patients infected by chronic hepatitis c and f0f1 fibrosis at baseline after a 15 years follow-up,” J. Hepatol., vol. 62, p. S589, Apr.2015.
[12] A. L. H.Øvrehus et al., “P0714 : Impact of prioritizing treatment in a high resource setting - minimizing the burden of HCV related disease in 15 years,” J. Hepatol., vol. 62, pp. S591–S592, Apr.2015.
[13] D.Petz, S.Klauck, F.-W.Röhl, P.Malfertheiner, A.Roessner, andC.Röcken, “Feasibility of histological grading and staging of chronic viral hepatitis using specimens obtained by thin-needle biopsy,” Virchows Arch., vol. 442, no. 3, pp. 238–244.
[14] G.Colloredo, M.Guido, A.Sonzogni, and G.Leandro, “Impact of liver biopsy size on histological evaluation of chronic viral hepatitis: the smaller the sample, the milder the disease,” J. Hepatol., vol. 39, no. 2, pp. 239–244, Aug.2003.
[15] F.Oberti et al., “Noninvasive diagnosis of hepatic fibrosis or cirrhosis,” Gastroenterology, vol. 113, no. 5, pp. 1609–1616, Nov.1997.
[16] F.Durand and D.Valla, “Assessment of the prognosis of cirrhosis: Child-Pugh versus MELD,” J. Hepatol., vol. 42, no. 1, pp. S100–S107, Apr.2005.
[17] H. S.Lee et al., “Prediction of compensated liver cirrhosis by ultrasonography and routine blood tests in patients with chronic viral hepatitis.,” Korean J. Hepatol., vol. 16, no. 4, pp. 369–75, Dec.2010.
[18] F.Salerno et al., “Prognostic value of the galactose test in predicting survival of patients with cirrhosis evaluated for liver transplantation: A prospective multicenter Italian study,” J. Hepatol., vol. 25, no. 4, pp. 474–480, Oct.1996.
[19] M.Zoli et al., “Prognostic significance of portal hemodynamics in patients with compensated cirrhosis,” J. Hepatol., vol. 17, no. 1, pp. 56–61, Jan.1993.
[20] C.Merkel et al., “Aminopyrine breath test in the prognostic evaluation of patients with cirrhosis.,” Gut, vol. 33, no. 6, pp. 836–42, Jun.1992.
[21] A.Colli, M.Fraquelli, M.Andreoletti, B.Marino, E.Zuccoli, and D.Conte, “Severe Liver Fibrosis or Cirrhosis: Accuracy of US for Detection—Analysis of 300 Cases,” Radiology, vol. 227, no. 1, pp. 89–94, Apr.2003.
[22] M.Viganò et al., “US Features of Liver Surface Nodularity as a Predictor of Severe Fibrosis in Chronic Hepatitis C,” Radiology, vol. 234, no. 2, pp. 641–641, Feb.2005.
[23] M.Soresi, L.Giannitrapani, M.Cervello, A.Licata, and G.Montalto, “Non invasive tools for the diagnosis of liver cirrhosis.,” World J. Gastroenterol., vol. 20, no. 48, pp. 18131–50, Dec.2014.
[24] R.Li, X.Hua, Y.Guo, P.Zhang, and A.Guo, “Neighborhood-pixels algorithm combined with Sono-CT in the diagnosis of cirrhosis: An experimental study,” Ultrasound Med. Biol., vol. 32, no. 10, pp. 1515–1520, Oct.2006.
[25] C.Vicas, M.Lupsor, M.Socaciu, S.Nedevschi, and R.Badea, “Influence of expert-dependent variability over the performance of noninvasive fibrosis assessment in patients with chronic hepatitis C by means of texture analysis.,” Comput. Math. Methods Med., vol. 2012, p. 346713, Dec.2012.
[26] S. K.Yeom, C. H.Lee, S. H.Cha, and C. M.Park, “Prediction of liver cirrhosis, using diagnostic imaging tools.,” World J. Hepatol., vol. 7, no. 17, pp. 2069–79, Aug.2015.
[27] M.Friedrich–Rust et al., “Performance of Transient Elastography for the Staging of Liver Fibrosis: A Meta-Analysis,” Gastroenterology, vol. 134, no. 4, pp. 960-974.e8, Apr.2008.
[28] I.Sporea et al., “Liver stiffness measurements in patients with HBV vs HCV chronic hepatitis: a comparative study.,” World J. Gastroenterol., vol. 16, no. 38, pp. 4832–7, Oct.2010.
[29] S.Riggio et al., “Assessment of liver stiffness in subjects affected by familial combined hyperlipidaemia with hepatic steatosis,” Eur. J. Clin. Invest., vol. 40, no. 8, pp. 722–728, Jun.2010.
[30] Y. E.Chon et al., “Performance of Transient Elastography for the Staging of Liver Fibrosis in Patients with Chronic Hepatitis B: A Meta-Analysis,” PLoS One, vol. 7, no. 9, p. e44930, Sep.2012.
[31] L.Annet, F.Peeters, J.Abarca-Quinones, I.Leclercq, P.Moulin, and B. E.VanBeers, “Assessment of diffusion-weighted MR imaging in liver fibrosis,” J. Magn. Reson. Imaging, vol. 25, no. 1, pp. 122–128, Jan.2007.
[32] D.LeBihan, E.Breton, D.Lallemand, M. L.Aubin, J.Vignaud, and M.Laval-Jeantet, “Separation of diffusion and perfusion in intravoxel incoherent motion MR imaging.,” Radiology, vol. 168, no. 2, pp. 497–505, Aug.1988.
[33] Y.Wang et al., “Assessment of Chronic Hepatitis and Fibrosis: Comparison of MR Elastography and Diffusion-Weighted Imaging,” Am. J. Roentgenol., vol. 196, no. 3, pp. 553–561, Mar.2011.
[34] L.-Q.Zou, J.Chen, L.Pan, J.-Z.Jiang, and W.Xing, “Comparison of magnetic resonance elastography and diffusion-weighted imaging for staging hepatic fibrosis.,” Chin. Med. J. (Engl)., vol. 128, no. 5, pp. 620–5, Mar.2015.
[35] R.Rustogi et al., “Accuracy of MR elastography and anatomic MR imaging features in the diagnosis of severe hepatic fibrosis and cirrhosis,” J. Magn. Reson. Imaging, vol. 35, no. 6, pp. 1356–1364, Jun.2012.
[36] S. K.Chamarthi et al., “Rapid acquisition technique for MR elastography of the liver,” Magn. Reson. Imaging, vol. 32, no. 6, pp. 679–683, Jul.2014.
[37] E. K.O’Neill, J. R.Cogley, and F. H.Miller, “The ins and outs of liver imaging.,” Clin. Liver Dis., vol. 19, no. 1, pp. 99–121, Feb.2015.
[38] M.Kudo et al., “Diagnostic Accuracy of Imaging for Liver Cirrhosis Compared to Histologically Proven Liver Cirrhosis,” Intervirology, vol. 51, no. 1, pp. 17–26, 2008.
[39] “Computer-aided diagnosis in medical imaging: Historical review, current status and future potential,” Comput. Med. Imaging Graph., vol. 31, no. 4–5, pp. 198–211, Jun.2007.
[40] U.Rajendra Acharya et al., “Data mining framework for fatty liver disease classification in ultrasound: A hybrid feature extraction paradigm,” 2012.
[41] A.Andrade, J. S.Silva, J.Santos, and P.Belo-Soares, “Classifier Approaches for Liver Steatosis using Ultrasound Images,” Procedia Technol., vol. 5, pp. 763–770, Jan.2012.
[42] M.Lupsor et al., “Non-invasive Steatosis Assessment through the Computerized Processing of Ultrasound Images: Attenuation versus First Order Texture Parameters,” Springer, Berlin, Heidelberg, 2011, pp. 184–189.
[43] F. ul A. A.Minhas, D.Sabih, and M.Hussain, “Automated Classification of Liver Disorders using Ultrasound Images,” J. Med. Syst., vol. 36, no. 5, pp. 3163–3172, Oct.2012.
[44] M.Singh, S.Singh, and S.Gupta, “An information fusion based method for liver classification using texture analysis of ultrasound images,” Inf. Fusion, vol. 19, pp. 91–96, Sep.2014.
[45] V. M.Gonçalves, M. E.Delamaro, and F. de L. dos S.Nunes, “A systematic review on the evaluation and characteristics of computer-aided diagnosis systems,” Rev. Bras. Eng. Biomédica, vol. 30, no. 4, pp. 355–383, Dec.2014.
[46] G.Castellano, L.Bonilha, L. M.Li, and F.Cendes, “Texture analysis of medical images,” Clin. Radiol., vol. 59, no. 12, pp. 1061–1069, Dec.2004.
[47] R. M.Haralick, K.Shanmugam, and I.Dinstein, “Textural Features for Image Classification,” IEEE Trans. Syst. Man. Cybern., vol. SMC-3, no. 6, pp. 610–621, Nov.1973.
[48] Xiaoou Tang, “Texture information in run-length matrices,” IEEE Trans. Image Process., vol. 7, no. 11, pp. 1602–1609, 1998.
[49] C.Sun and W. G.Wee, “Neighboring gray level dependence matrix for texture classification,” Comput. Vision, Graph. Image Process., vol. 23, no. 3, pp. 341–352, Sep.1983.
[50] M.-H.Horng, Y.-N.Sun, andX.-Z.Lin, “Texture feature coding method for classification of liver sonography,” Comput. Med. Imaging Graph., vol. 26, no. 1, pp. 33–42, Jan.2002.
[51] T.Mitchell, B.Buchanan, G.Dejong, T.Dietterich, P.Rosenbloom, and A.Waibel, “MACHINE LEARNING,” 1990.
[52] K. P.Bennett and C.Campbell, “Support vector machines,” ACM SIGKDD Explor. Newsl., vol. 2, no. 2, pp. 1–13, Dec.2000.
[53] G. P.Zhang, “Neural networks for classification: a survey,” IEEE Trans. Syst. Man Cybern. Part C (Applications Rev., vol. 30, no. 4, pp. 451–462, 2000.
[54] N.Lavesson and P.Davidsson, “Generic Methods for Multi-criteria Evaluation,” in Proceedings of the 2008 SIAM International Conference on Data Mining, 2008, pp. 541–546.
[55] N. H.Sweilam, A. A.Tharwat, and N. K.Abdel Moniem, “Support vector machine for diagnosis cancer disease: A comparative study,” Egypt. Informatics J., vol. 11, no. 2, pp. 81–92, Dec.2010.
[56] M.Mohaimenul Islam, C.-C.Wu, T.Nasrin Poly, H.-C.Yang, M.Islam, andY.-C.Li, “Applications of Machine Learning in Fatty Live Disease Prediction,” 2018.
[57] O. S.Soliman and A.Elhamd, “Classification of Hepatitis C Virus using Modified Particle Swarm Optimization and Least Squares Support Vector Machine,” Int. J. Sci. Eng. Res., vol. 5, no. 3, 2014.
[58] N.Nahar and F.Ara, “LIVER DISEASE PREDICTION BY USING DIFFERENT DECISION TREE TECHNIQUES,” Int. J. Data Min. Knowl. Manag. Process, vol. 8, no. 2, 2018.
[59] R.Wei et al., “Clinical prediction of HBV and HCV related hepatic fibrosis using machine learning,” EBioMedicine, vol. 35, pp. 124–132, Sep.2018.
[60] E.Grossi and M.Buscema, “Introduction to artificial neural networks,” Eur. J. Gastroenterol. Hepatol., vol. 19, no. 12, pp. 1046–1054, Dec.2007.
[61] S.Sontakke, J.Lohokare, and R.Dani, “Diagnosis of liver diseases using machine learning,” in 2017 International Conference on Emerging Trends & Innovation in ICT (ICEI), 2017, pp. 129–133.
[62] G. E.Hinton, S.Osindero, andY.-W.Teh, “A Fast Learning Algorithm for Deep Belief Nets,” Neural Comput., vol. 18, no. 7, pp. 1527–1554, Jul.2006.
[63] Y.Bengio, P.Lamblin, D.Popovici, and H.Larochelle, “Greedy Layer-Wise Training of Deep Networks.”
[64] L.-J.Li et al., “ImageNet: a Large-Scale Hierarchical Image Database,” 2009.
[65] M. Z.Alom et al., “The History Began from AlexNet: A Comprehensive Survey on Deep Learning Approaches,” Mar.2018.
[66] “Deep Learning in Five and a Half Minutes.” [Online]. Available: https://www.embedded-vision.com/industry-analysis/blog/deep-learning-five-and-half-minutes. [Accessed: 08-Jun-2019].
[67] K.Simonyan and A.Zisserman, “Very Deep Convolutional Networks for Large-Scale Image Recognition,” Sep.2014.
[68] C.Szegedy et al., “Going Deeper With Convolutions.” pp. 1–9, 2015.
[69] K.He, X.Zhang, S.Ren, andJ.Sun, “Deep Residual Learning for Image Recognition.” pp. 770–778, 2016.
[70] A.Esteva et al., “Dermatologist-level classification of skin cancer with deep neural networks,” Nature, vol. 542, no. 7639, pp. 115–118, Feb.2017.
[71] V.Gulshan et al., “Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs,” JAMA, vol. 316, no. 22, p. 2402, Dec.2016.
[72] E.Alberdi, A. A.Povyakalo, L.Strigini, and P.Ayton, “Computer Aided Detection: Risks and benefits for radiologists’ decisions.”
[73] J. J.Cerrolaza, M.Reyes, R. M.Summers, M. Á.González-Ballester, and M. G.Linguraru, “Automatic multi-resolution shape modeling of multi-organ structures,” Med. Image Anal., vol. 25, no. 1, pp. 11–21, Oct.2015.
[74] T.Okada, M. G.Linguraru, M.Hori, R. M.Summers, N.Tomiyama, andY.Sato, “Abdominal multi-organ segmentation from CT images using conditional shape–location and unsupervised intensity priors,” Med. Image Anal., vol. 26, no. 1, pp. 1–18, Dec.2015.
[75] T.Rohlfing, R.Brandt, R.Menzel, and C. R.Maurer, “Evaluation of atlas selection strategies for atlas-based image segmentation with application to confocal microscopy images of bee brains,” Neuroimage, vol. 21, no. 4, pp. 1428–1442, Apr.2004.
[76] H.Wang et al., “Multi-atlas Segmentation with Robust Label Transfer and Label Fusion,” Springer, Berlin, Heidelberg, 2013, pp. 548–559.
[77] J. E.Iglesias and M. R.Sabuncu, “Multi-atlas segmentation of biomedical images: A survey,” Med. Image Anal., vol. 24, no. 1, pp. 205–219, Aug.2015.
[78] T.Tong et al., “Discriminative dictionary learning for abdominal multi-organ segmentation,” Med. Image Anal., vol. 23, no. 1, pp. 92–104, Jul.2015.
[79] M.Oda et al., “Regression Forest-Based Atlas Localization and Direction Specific Atlas Generation for Pancreas Segmentation,” Springer, Cham, 2016, pp. 556–563.
[80] K.Kamnitsas et al., “Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation,” Med. Image Anal., vol. 36, pp. 61–78, Feb.2017.
[81] X.Zhou, T.Ito, R.Takayama, S.Wang, T.Hara, and H.Fujita, “Three-Dimensional CT Image Segmentation by Combining 2D Fully Convolutional Network with 3D Majority Voting,” Springer, Cham, 2016, pp. 111–120.
[82] F.Milletari, N.Navab, andS.-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.
[83] F.Albregtsen, “Statistical Texture Measures Computed from Gray-Level Co-occurrence Matrices,” 2008.
[84] Xiaoou Tang, “Texture information in run-length matrices,” IEEE Trans. Image Process., vol. 7, no. 11, pp. 1602–1609, 1998.
[85] C.Sun and W. G.Wee, “Neighboring gray level dependence matrix for texture classification,” Comput. Vision, Graph. Image Process., vol. 23, no. 3, pp. 341–352, Sep.1983.
[86] I.Mukherjee and S.Routroy, “Comparing the performance of neural networks developed by using Levenberg–Marquardt and Quasi-Newton with the gradient descent algorithm for modeling a multiple response grinding process,” Expert Syst. Appl., vol. 39, no. 3, pp. 2397–2407, Feb.2012.
[87] “3Dircadb | IRCAD France.” [Online]. Available: https://www.ircad.fr/research/3dircadb/. [Accessed: 18-Jun-2019].
[88] S. S. M.Salehi, D.Erdogmus, and A.Gholipour, “Tversky Loss Function for Image Segmentation Using 3D Fully Convolutional Deep Networks,” Springer, Cham, 2017, pp. 379–387.