深度学习技术在胃肠镜诊断中应用研究进展
2020-07-04姜良慧周长宏
姜良慧 周长宏
[摘要]深度学习(DL)为目前主流的人工智能图像识别方法,在医疗领域受到越来越多关注。DL技术通过对海量训练数据中的原始特征进行表征分析,构建多隐层的机器学习模型,模拟脑神经元活动,最后计算机输出单一的诊断结果。本文结合国内外研究成果,介绍DL在Barrett食管、食管癌、幽门螺旋杆菌感染、胃癌、结肠息肉等方面的应用,以及DL在胃肠镜诊断中存在的不足及未来发展前景。
[关键词] 人工智能;机器学习;诊断,计算机辅助;胃镜检查;结肠镜检查;综述
[中图分类号] R319
[文献标志码] A
[文章编号] 2096-5532(2020)03-0373-06
doi:10.11712/jms.2096-5532.2020.56.076
[开放科学(资源服务)标识码(OSID)]
[网络出版] http://kns.cnki.net/kcms/detail/37.1517.R.20200417.0906.002.html;2020-04-18 15:26
RESEARCH ADVANCES IN THE APPLICATION OF DEEP LEARNING TECHNOLOGY IN DIAGNOSIS USING GASTROINTESTINAL ENDOSCOPY
JIANG Lianghui, ZHOU Changhong
(Health Care Section, The Affiliated Qingdao Municipal Hospital of Qingdao University, Qingdao 266003, China)
[ABSTRACT]
Deep learning, as the mainstream artificial intelligence image recognition method, has attracted more and more attention in the medical field. Through a characterization analysis of the original features of massive training data, deep lear-
ning technology can build a machine learning model with multiple hidden layers, simulate the activity of cerebral neurons, and finally give a single diagnostic result via computer. With reference to related research findings in China and foreign countries, this article introduces the application of deep learning technology in Barretts esophagus, esophageal cancer, Helicobacter pylori infection, gastric cancer, and colonic polyps, as well as its shortcomings and
future prospects in diagnosis with gastrointestinal endoscopy.
[KEY WORDS]artificial intelligence; machine learning; diagnosis, computer-assisted; gastroscopy; colonoscopy; review
消化内镜技术是诊断胃肠道疾病的重要手段,普通白光内镜在观察黏膜浅表血管、组织形态变化方面已无法满足临床需求。随着窄带成像放大内镜、智能分光比色技术、共聚焦激光显微内镜等较为先进内镜技术应用于临床,内镜成像包含越来越多的复杂信息,内镜图像的解读成为一项具有挑战性工作。随着信息技术快速发展,深度学习(DL)在医疗领域受到越来越多的关注,计算机辅助诊断(CAD)胃肠道疾病成为一项热门的研究课题。通过分析提取疾病信息载体的内镜图像数据特征,CAD可提高临床医生执行医疗任务能力,并有效提升诊断准确率,进而指导治疗、评估预后。目前,DL已在内镜下诊断Barrett食管、食管癌、幽门螺旋杆菌(Hp)感染、胃癌、结肠息肉等方面取得一定成果,未来在胃肠病学中应用范围将进一步扩大。……
