近年來，非酒精性脂肪性肝病（NAFLD）已成為全球健康關切的重要議題。NAFLD的主要病理特徵，脂肪肝，傳統上由臨床醫師或實驗室技術人員進行評分以評估疾病的嚴重程度。人工智能技術預計將通過提供更快速且定量的組織中脂肪積累評估，有助於診斷。然而，當前的人工智能模型在準確分割和描述NAFLD動物模型中脂肪積累的程度方面仍存在不足。我們提出了一種稱為基於課程學習的全感知脂肪肝分割（CuFASS）算法，該算法將邊界感知模型與課程學習技術相結合，使模型能夠學習全面的脂肪肝形態。我們對1722張數位影像，來自22張全切片影像蘇木精-伊紅染色（H＆E）標本的病理影像進行了評估，根據脂肪肝的嚴重程度進行分類。分析結果顯示，我們的模型在NAFLD組中實現了81.3％的精確度與召回率的調和平均數（F1-score），85.0％的精確度和79.9％的召回率。此外，模型的預測結果與臨床醫師計算的半定量分數進行了比較，通過相關性分析計算出了0.834的決定係數值。我們的模型在檢測小鼠脂肪肝疾病中泡狀脂肪積累方面表現出良好的精確度和敏感度。

Non-alcoholic fatty liver disease (NAFLD) has become a significant global health concern in recent times. Steatosis, the major pathological feature of NAFLD, is traditionally scored by clinicians or laboratory technicians to evaluate the severity of the disease. AI technologies are expected to aid in diagnosis by offering a faster and quantitative assay of lipid accumulation in tissue. However, the current AI models have been inadequate in accurately segmenting and delineating the extent of lipid accumulation in the NAFLD animal model. We propose a method designated as Curriculum Learning-based Fully-aware Steatosis Segmentation (CuFASS) algorithm that integrates a boundary-aware model with curriculum learning techniques, enabling the model to learn comprehensive steatosis shapes. 1722 digital images from 22 Hematoxylin and Eosin (H&E) stained whole slide images (WSIs) were categorized by the severity of steatosis. The analysis indicated our present model achieved a 81.3% F1-score, 85.0% precision, and 79.9% recall in NAFLD group. Moreover, the model's predictions were compared with the semi-quantitative scores computed by clinicians, yielding a coefficient of determination value of 0.834 through correlation analysis. Our model demonstrates performance in terms of precision and sensitivity for detecting vesicular steatosis in mouse fatty liver disease.

[1] “Global epidemiology of NAFLD-related HCC: trends, predictions, risk factors and prevention - PubMed.” https://pubmed.ncbi.nlm.nih.gov/33349658/ (accessed Jun. 18, 2023).
[2] C. Estes, H. Razavi, R. Loomba, Z. Younossi, and A. J. Sanyal, “Modeling the epidemic of nonalcoholic fatty liver disease demonstrates an exponential increase in burden of disease,” Hepatology, vol. 67, no. 1, pp. 123–133, Jan. 2018, doi: 10.1002/hep.29466.
[3] A. C. Sheka, O. Adeyi, J. Thompson, B. Hameed, P. A. Crawford, and S. Ikramuddin, “Nonalcoholic Steatohepatitis: A Review,” JAMA, vol. 323, no. 12, pp. 1175–1183, Mar. 2020, doi: 10.1001/jama.2020.2298.
[4] Z. M. Younossi, A. B. Koenig, D. Abdelatif, Y. Fazel, L. Henry, and M. Wymer, “Global epidemiology of nonalcoholic fatty liver disease-Meta-analytic assessment of prevalence, incidence, and outcomes,” Hepatology, vol. 64, no. 1, pp. 73–84, Jul. 2016, doi: 10.1002/hep.28431.
[5] D. E. Kleiner et al., “Design and validation of a histological scoring system for nonalcoholic fatty liver disease,” Hepatology, vol. 41, no. 6, pp. 1313–1321, Jun. 2005, doi: 10.1002/hep.20701.
[6] F. Sapmaz et al., “Steatosis Grade is the Most Important Risk Factor for Development of Endothelial Dysfunction in NAFLD,” Medicine (Baltimore), vol. 95, no. 14, p. e3280, Apr. 2016, doi: 10.1097/MD.0000000000003280.
[7] N. Chalasani et al., “Relationship of steatosis grade and zonal location to histological features of steatohepatitis in adult patients with non-alcoholic fatty liver disease,” J Hepatol, vol. 48, no. 5, pp. 829–834, May 2008, doi: 10.1016/j.jhep.2008.01.016.
[8] A. J. Sanyal et al., “Endpoints and clinical trial design for nonalcoholic steatohepatitis,” Hepatology, vol. 54, no. 1, pp. 344–353, Jul. 2011, doi: 10.1002/hep.24376.
[9] N. Selzner, M. Selzner, W. Jochum, B. Amann-Vesti, R. Graf, and P.-A. Clavien, “Mouse livers with macrosteatosis are more susceptible to normothermic ischemic injury than those with microsteatosis,” J Hepatol, vol. 44, no. 4, pp. 694–701, Apr. 2006, doi: 10.1016/j.jhep.2005.07.032.
[10] D. David and C. E. Eapen, “What Are the Current Pharmacological Therapies for Nonalcoholic Fatty Liver Disease?,” J Clin Exp Hepatol, vol. 11, no. 2, pp. 232–238, 2021, doi: 10.1016/j.jceh.2020.09.001.
[11] G. Quintás, J. V. Castell, and M. Moreno-Torres, “The assessment of the potential hepatotoxicity of new drugs by in vitro metabolomics,” Front Pharmacol, vol. 14, p. 1155271, May 2023, doi: 10.3389/fphar.2023.1155271.
[12] D. Altunbay, C. Cigir, C. Sokmensuer, and C. Gunduz-Demir, “Color Graphs for Automated Cancer Diagnosis and Grading,” IEEE Transactions on Biomedical Engineering, vol. 57, no. 3, pp. 665–674, Mar. 2010, doi: 10.1109/TBME.2009.2033804.
[13] N. Batool, “Detection and spatial analysis of hepatic steatosis in histopathology images using sparse linear models,” in 2016 Sixth International Conference on Image Processing Theory, Tools and Applications (IPTA), Feb. 2016, pp. 1–6. doi: 10.1109/IPTA.2016.7820969.
[14] M. De Rudder et al., “Automated computerized image analysis for the user-independent evaluation of disease severity in preclinical models of NAFLD/NASH,” Lab Invest, vol. 100, no. 1, Art. no. 1, Jan. 2020, doi: 10.1038/s41374-019-0315-9.
[15] M. Salvi, N. Michielli, and F. Molinari, “Stain Color Adaptive Normalization (SCAN) algorithm: Separation and standardization of histological stains in digital pathology,” Computer Methods and Programs in Biomedicine, vol. 193, p. 105506, Sep. 2020, doi: 10.1016/j.cmpb.2020.105506.
[16] Y. Ramot, G. Zandani, Z. Madar, S. Deshmukh, and A. Nyska, “Utilization of a Deep Learning Algorithm for Microscope-Based Fatty Vacuole Quantification in a Fatty Liver Model in Mice,” Toxicol Pathol, vol. 48, no. 5, pp. 702–707, Jul. 2020, doi: 10.1177/0192623320926478.
[17] M. Roy et al., “Deep-learning-based accurate hepatic steatosis quantification for histological assessment of liver biopsies,” Lab Invest, vol. 100, no. 10, Art. no. 10, Oct. 2020, doi: 10.1038/s41374-020-0463-y.
[18] R. Forlano et al., “High-Throughput, Machine Learning-Based Quantification of Steatosis, Inflammation, Ballooning, and Fibrosis in Biopsies From Patients With Nonalcoholic Fatty Liver Disease,” Clin Gastroenterol Hepatol, vol. 18, no. 9, pp. 2081-2090.e9, Aug. 2020, doi: 10.1016/j.cgh.2019.12.025.
[19] D. Sethunath et al., “Automated assessment of steatosis in murine fatty liver,” PLoS One, vol. 13, no. 5, p. e0197242, 2018, doi: 10.1371/journal.pone.0197242.
[20] “ResUNet++: An Advanced Architecture for Medical Image Segmentation | IEEE Conference Publication | IEEE Xplore.” https://ieeexplore.ieee.org/document/8959021 (accessed Jul. 02, 2023).
[21] X. Wei, X. Gong, Y. Zhan, B. Du, Y. Luo, and W. Hu, “CLNode: Curriculum Learning for Node Classification,” in Proceedings of the Sixteenth ACM International Conference on Web Search and Data Mining, in WSDM ’23. New York, NY, USA: Association for Computing Machinery, Feb. 2023, pp. 670–678. doi: 10.1145/3539597.3570385.
[22] J. Wei et al., “Learn Like a Pathologist: Curriculum Learning by Annotator Agreement for Histopathology Image Classification,” presented at the Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021, pp. 2473–2483. Accessed: Jun. 19, 2023. [Online]. Available: https://openaccess.thecvf.com/content/WACV2021/html/Wei_Learn_Like_a_Pathologist_Curriculum_Learning_by_Annotator_Agreement_for_WACV_2021_paper.html
[23] “Medical knowledge-guided deep curriculum learning for elbow fracture diagnosis from x-ray images.” https://www.spiedigitallibrary.org/conference-proceedings-of-spie/11597/1159712/Medical-knowledge-guided-deep-curriculum-learning-for-elbow-fracture-diagnosis/10.1117/12.2582184.full?webSyncID=d883c9d9-02bc-9993-ced2-68bead49a285&sessionGUID=31ab0d3a-6b03-c97c-009b-6bcd746b80d8&SSO=1 (accessed Jul. 03, 2023).
[24] Y. Bengio, J. Louradour, R. Collobert, and J. Weston, “Curriculum learning,” in Proceedings of the 26th Annual International Conference on Machine Learning, in ICML ’09. New York, NY, USA: Association for Computing Machinery, Jun. 2009, pp. 41–48. doi: 10.1145/1553374.1553380.
[25] “Theory of curriculum learning, with convex loss functions | The Journal of Machine Learning Research.” https://dl.acm.org/doi/abs/10.5555/3455716.3455938 (accessed Aug. 22, 2023).
[26] D. Weinshall, G. Cohen, and D. Amir, “Curriculum Learning by Transfer Learning: Theory and Experiments with Deep Networks,” in Proceedings of the 35th International Conference on Machine Learning, PMLR, Jul. 2018, pp. 5238–5246. Accessed: Aug. 22, 2023. [Online]. Available: https://proceedings.mlr.press/v80/weinshall18a.html
[27] G. Hacohen and D. Weinshall, “On The Power of Curriculum Learning in Training Deep Networks,” in Proceedings of the 36th International Conference on Machine Learning, PMLR, May 2019, pp. 2535–2544. Accessed: Jul. 17, 2023. [Online]. Available: https://proceedings.mlr.press/v97/hacohen19a.html
[28] Y. Huang et al., “CurricularFace: Adaptive Curriculum Learning Loss for Deep Face Recognition,” in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, USA: IEEE, Jun. 2020, pp. 5900–5909. doi: 10.1109/CVPR42600.2020.00594.
[29] B. Zhang et al., “FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labeling,” in Advances in Neural Information Processing Systems, Curran Associates, Inc., 2021, pp. 18408–18419. Accessed: Aug. 22, 2023. [Online]. Available: https://proceedings.neurips.cc/paper/2021/hash/995693c15f439e3d189b06e89d145dd5-Abstract.html
[30] M. de Lhoneux, S. Zhang, and A. Søgaard, “Zero-Shot Dependency Parsing with Worst-Case Aware Automated Curriculum Learning,” in Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), Dublin, Ireland: Association for Computational Linguistics, May 2022, pp. 578–587. doi: 10.18653/v1/2022.acl-short.64.
[31] C. Wang, Y. Wu, S. Liu, M. Zhou, and Z. Yang, “Curriculum Pre-training for End-to-End Speech Translation.” arXiv, Apr. 21, 2020. doi: 10.48550/arXiv.2004.10093.
[32] J. L. Elman, “Learning and development in neural networks: the importance of starting small,” Cognition, vol. 48, no. 1, pp. 71–99, Jul. 1993, doi: 10.1016/0010-0277(93)90058-4.
[33] L. Jiang, D. Meng, S.-I. Yu, Z. Lan, S. Shan, and A. Hauptmann, “Self-Paced Learning with Diversity,” in Advances in Neural Information Processing Systems, Curran Associates, Inc., 2014. Accessed: Aug. 22, 2023. [Online]. Available: https://proceedings.neurips.cc/paper/2014/hash/c60d060b946d6dd6145dcbad5c4ccf6f-Abstract.html
[34] L. Jiang, D. Meng, T. Mitamura, and A. G. Hauptmann, “Easy Samples First: Self-paced Reranking for Zero-Example Multimedia Search,” in Proceedings of the 22nd ACM international conference on Multimedia, in MM ’14. New York, NY, USA: Association for Computing Machinery, Nov. 2014, pp. 547–556. doi: 10.1145/2647868.2654918.
[35] S. Gururangan et al., “Don’t Stop Pretraining: Adapt Language Models to Domains and Tasks,” in Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, Online: Association for Computational Linguistics, Jul. 2020, pp. 8342–8360. doi: 10.18653/v1/2020.acl-main.740.
[36] T. Karras, T. Aila, S. Laine, and J. Lehtinen, “Progressive Growing of GANs for Improved Quality, Stability, and Variation.” arXiv, Feb. 26, 2018. doi: 10.48550/arXiv.1710.10196.
[37] A. Graves, A. Mohamed, and G. Hinton, “Speech recognition with deep recurrent neural networks,” in 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, May 2013, pp. 6645–6649. doi: 10.1109/ICASSP.2013.6638947.
[38] R. E. Wray, J. R. Kirk, and J. E. Laird, “Language Models as a Knowledge Source for Cognitive Agents”.
[39] C. Matuszek, N. FitzGerald, L. Zettlemoyer, L. Bo, and D. Fox, “A Joint Model of Language and Perception for Grounded Attribute Learning.” arXiv, Jun. 27, 2012. doi: 10.48550/arXiv.1206.6423.
[40] M. Ren, W. Zeng, B. Yang, and R. Urtasun, “Learning to Reweight Examples for Robust Deep Learning,” in Proceedings of the 35th International Conference on Machine Learning, PMLR, Jul. 2018, pp. 4334–4343. Accessed: Aug. 22, 2023. [Online]. Available: https://proceedings.mlr.press/v80/ren18a.html
[41] M. Salvi et al., “Fully automated quantitative assessment of hepatic steatosis in liver transplants,” Computers in Biology and Medicine, vol. 123, p. 103836, Aug. 2020, doi: 10.1016/j.compbiomed.2020.103836.
[42] D. D. Mais, Practical Clinical Pathology. American Society for Clinical Pathology, 2013.
[43] J.-S. Wu, “Region and Boundary Probability-based Instance Segmentation for Lipid Droplet Quantification in Diagnosis of Non-Alcoholic Fatty Liver Disease,” National Cheng Kung University, 2021.
[44] S. S. M. Salehi, D. Erdogmus, and A. Gholipour, “Tversky Loss Function for Image Segmentation Using 3D Fully Convolutional Deep Networks,” in Machine Learning in Medical Imaging, Q. Wang, Y. Shi, H.-I. Suk, and K. Suzuki, Eds., in Lecture Notes in Computer Science. Cham: Springer International Publishing, 2017, pp. 379–387. doi: 10.1007/978-3-319-67389-9_44.
[45] C. C. Chinglemba and P. Chungkham, “Study of Loss Functions on Retinal Vessel Segmentation using UNET Architecture,” in 2021 International Conference on Disruptive Technologies for Multi-Disciplinary Research and Applications (CENTCON), Jan. 2021, pp. 308–312. doi: 10.1109/CENTCON52345.2021.9688289.
[46] X. Chen, B. M. Williams, S. R. Vallabhaneni, G. Czanner, R. Williams, and Y. Zheng, “Learning Active Contour Models for Medical Image Segmentation,” in 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2019, pp. 11624–11632. doi: 10.1109/CVPR.2019.01190.
[47] A. Jiménez-Sánchez et al., “Medical-based Deep Curriculum Learning for Improved Fracture Classification,” in Medical Image Computing and Computer Assisted Intervention – MICCAI 2019, D. Shen, T. Liu, T. M. Peters, L. H. Staib, C. Essert, S. Zhou, P.-T. Yap, and A. Khan, Eds., in Lecture Notes in Computer Science. Cham: Springer International Publishing, 2019, pp. 694–702. doi: 10.1007/978-3-030-32226-7_77.
[48] C. Agarwal, D. D’souza, and S. Hooker, “Estimating Example Difficulty using Variance of Gradients,” in 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Jun. 2022, pp. 10358–10368. doi: 10.1109/CVPR52688.2022.01012.
[49] Y. Kong, L. Liu, J. Wang, and D. Tao, “Adaptive Curriculum Learning,” presented at the Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 5067–5076. Accessed: Jul. 02, 2023. [Online]. Available: https://openaccess.thecvf.com/content/ICCV2021/html/Kong_Adaptive_Curriculum_Learning_ICCV_2021_paper.html
[50] O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation,” in Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015, N. Navab, J. Hornegger, W. M. Wells, and A. F. Frangi, Eds., in Lecture Notes in Computer Science. Cham: Springer International Publishing, 2015, pp. 234–241. doi: 10.1007/978-3-319-24574-4_28.