您当前的位置:首页 >> 用图
CULTURE丨新型人工智能模型或帮助解决近视问题
发布时间:2019-07-30
 

AI model helps solve myopia problems



By                          / Ye Jiao

Proofreader / Zhao Siyuan

Photo                   / Internet


After analyzing millions of medical optometry big data, Liu Yizhi's team from Zhongshan Ophthalmic Center of Sun Yat-sen University created an artificial intelligence prediction model for myopia, which is now applied to a cloud platform published on November 8, to provide myopia variation tendency prediction services individually.

在分析了数以百万计的医疗验光数据后,来自中山大学中山眼科中心的刘奕志及他的团队创建了预测近视变化趋势的人工智能模型,并将其应用在11月8日发布的云平台上,为个人提供预测服务。



In the decade from 2005 to 2015, the research team collected 1.25 million optometric results from eight medical institutions. “All these numerous data helped us find rules in the development of myopia in adolescents.” Said Liu.

2005年至2015年的十年间,该研究团队从八家医疗机构收集了一百二十五万余份验光结果。“所有这些大量的数据帮助我们找到了青少年近视发展的规律。”刘奕志说。


By accessing the smart platform and inputting the age and degree of the two myopia records taken at least one-year interval, the myopia degree changes within 10 years and the risk of high myopia before age 18 can be predicted.

用户可以通过访问智能平台,输入间隔至少一年的两次近视诊察记录的年龄和近视度数,来预测十年内近视度数的变化情况以及十八岁前患高度近视的风险大小。


The diagnostic accuracy is 90 percent within three years, and 80 percent within 10 years. Thus, by means of implementing early intervention such as increasing outdoor activities can slow down the aggravation of myopia and reduce the risk of high myopia.

预测未来三年内变化情况的诊断准确率能达到90%,未来10年的准确率则达到了80%。如此,通过对患者进行早期干预,如增加患者的户外活动,可以减缓患者近视的加重,降低高度近视的风险。


Type setting: Zhang Chi

Instructors: Zhang Pei & Chen Yu

Chief editor: Yu Jinsong

New media director: Zhang Chi