Home > News content

World's first: Sun Yat-sen University has developed a mobile phone intelligent screening system for infants with visual impairment

via:IT之家     time:2023/1/31 14:01:44     readed:326

Home of IT January 31 news, according to the official public account of Sun Yat-sen University news, Sun Yat-sen Ophthalmic Center of Sun Yat-sen University held a press conference on January 30, announcing that the team of Professor Lin Hao-tian, deputy director of Sun Yat-sen Ophthalmic Center, and a number of institutions around the world worked together to research and develop the world's first intelligent mobile phone screening system for infants with visual impairment. Parents can use smartphones to test their children for 16 common eye diseases that cause blindness. The results were published online January 26 in the international academic journal Nature Medicine.

图片

The paper is published online in the journal Nature Medicine

Professor Yu Minbin, deputy secretary of the Party Committee of Sun Yat-sen University and a renowned ophthalmologist, said that infants aged 0 to 4 years old are at a critical stage of visual function development, so it is important to assess infants' visual function and early detection of visual damage caused by eye diseases. However, it is difficult for infants in this age group to express eye discomfort, and it is difficult to cooperate with the traditional eye examination. Their visual function injury is easy to be ignored or missed, and the best treatment opportunity is missed.

The research and development team found in clinical practice that some abnormal fixation behavior patterns in infants were highly correlated with visual impairment. In view of this, the research and development team spent 8 years to prospectively collect the big data of staring behavior from 3652 infants. Relying on the deep learning artificial intelligence data analysis background deployed by Tianhe-2 supercomputing Center, the collected high-throughput video data was deeply analyzed. Moreover, based on big data, the abnormal fixation behavior modes of infants' visual function impairment caused by different eye diseases were deeply mined, and artificial intelligence models were constructed to realize the early detection of visual function impairment caused by 16 common blindness diseases in infants, such as congenital cataract, congenital ptosis and congenital glaucoma, with an average screening accuracy of over 85%.

图片

The use of smart phones for early screening of infant visual impairment scenes

图片

Smart phones play animated videos and record infants' gaze behavior

Lin Haotian, deputy director of the Zhongshan Ophthalmic Center at Sun Yat-sen University and lead author of the study, said the system can capture infants' staring habits and behavior patterns in real time by playing a three-minute animated video on a mobile phone to attract them to keep watching. Subsequently, the system relies on artificial intelligence technology to automatically analyze infant visual damage and related eye diseases.

It is reported that the intelligent screening system has a stable performance in complex realistic screening scenarios, which is suitable for hospital, community, home and other scenarios. It will help the early screening of a variety of infant eye diseases related to visual function impairment, greatly reducing the difficulty and cost.

Currently, members of the public can try out the screening system by searching for "Ai Baofingling" in the wechat mini program, IT Home has learned. The next small program "Yuejingjing" also plans to introduce the screening system. In addition, the research and development team is working with public platforms such as Guangdong Provincial Affairs, hoping to provide feasible programs for the mass screening of infants with visual function injury and related eye diseases.

图片

translate engine: Youdao

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments