Yao B-Y, Tian Z, Wu H-Y, Ma L-M
Department of General Surgery, Shengjing Hospital, China Medical University, Shenyang, Liaoning, China.
Eur Rev Med Pharmacol Sci. 2021 Mar;25(5):2245-2251. doi: 10.26355/eurrev_202103_25256.
As we know, gallstones are a gallbladder disease with high incidence around the world. As the population has aged and living habits have changed, the incidence of the disease is increasing year by year. Gallstones are mainly classified into cholesterol, bile pigment and mixed type gallstones based on their chemical composition. Patients with different stone components have different treatment options. Therefore, it is very important to know the chemical type of the stone before treatment. Imaging examination is the main method to identify the components of gallstones in the body.
Deep learning technology has an excellent data mining ability, and thus the combination of deep learning and medical treatment is always a research focus. In this work, we introduce a generative model to learn the features of the training data, to detect the composition of gallstones and to assist medical diagnosis. Furthermore, the theoretical analysis is given in detail.
The model could be used to determine the chemical composition of gallstones.
The potential of generative models in predicting the chemical composition of gallstones is shown in this study. In addition, theoretical analysis is also presented.
众所周知，胆结石是一种在全球发病率很高的胆囊疾病。随着人口老龄化和生活习惯的改变，该疾病的发病率逐年上升。胆结石根据其化学成分主要分为胆固醇结石、胆色素结石和混合型结石。不同结石成分的患者有不同的治疗方案。因此，在治疗前了解结石的化学类型非常重要。影像学检查是识别体内胆结石成分的主要方法。
深度学习技术具有出色的数据挖掘能力，因此深度学习与医学治疗的结合一直是研究热点。在这项工作中，我们引入了一种生成模型来学习训练数据的特征，检测胆结石的成分并辅助医学诊断。此外，还给出了详细的理论分析。
该模型可用于确定胆结石的化学成分。
本研究展示了生成模型在预测胆结石化学成分方面的潜力。此外，还进行了理论分析。
Zotero 插件