Exploring the diagnosis and syndrome differentiation of chronic gastritis with Chinese and Western medicine based on machine learning and gastroscopy
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
目的 探索多种机器学习模型基于胃镜下胃黏膜状态预测慢性胃炎(CG)的分类诊断与分型辨证,为中西医诊断理论创新提供价值。 方法 回顾性收集诊断为慢性胃炎的2052张胃镜图像,进行数据预处理后,建立3种机器学习模型,评估机器学习算法辅助下胃镜在慢性胃炎诊断中的价值。胃镜作为中医望诊的延伸,在原有数据中筛选2016张胃镜图像作为新的数据集,基于机器学习进行虚实辨证,探索胃镜在慢性胃炎中医辨证分型中的作用。 结果 训练的3种基于胃镜的模型(ViT、ResNet50、VGG16)对慢性萎缩性胃炎(CAG)的准确率分别98.60%、95.16%、95.81%,识别慢性非萎缩性胃炎(CNAG)准确率依次为99.16%、99.58%、99.16%;进一步实验的4种机器学习模型(ViT、ResNet50、VGG16、Mobilenet)识别CG实证组准确率依次为95.09%、88.24%、90.20%、87.25%;CG虚证组准确率为:89.10%、88.11%、89.14%、91.09%。 结论 机器学习可以作为辅助CG胃镜下诊断的稳定方法,且将课题组开发的基于胃镜的机器学习诊断模型可用于CG的中医虚实辨证。机器学习可辅助医师对该病进行识分类、辨证等任务,为人机共同决策提供临床应用价值。
Fields of Study & Subjects
AI in cancer detection
Artificial Intelligence in Healthcare
Helicobacter pylori-related gastroenterology studies
Publication details
- Journal / source
- Lishizhen Medicine and Materia Medica Research
- Volume
- 37
- Issue
- 5
- Pages
- 993–1000
- Publication date
- 1 March 2026
- Language
- EN