• 作者：Hermanus C. Myburgh^a ; Willemien H. van Zijl^a ; DeWet Swanepoel^b ; ^c ; ^g ; Sten Hellströ ; m^d ; ^e ; Claude Laurent^b ; ^f ; ^{claude.laurent@umu.se" class="auth_mail" title="E-mail the corresponding author}
• 关键词：Decision tree ; Feature extraction ; Automated diagnosis ; Otitis media ; Video-otoscope ; Global medicine
• 刊名：EBioMedicine
• 出版年：2016
• 出版时间：March 2016
• 年：2016
• 卷：5
• 期：Complete
• 页码：156-160
• 全文大小：504 K

•

Computer-based otitis media classification can provide access to diagnosis in developing countries.

•

Diagnostic accuracy of the image-analysis classification system was comparable to general practitioners' and pediatricians'.

Most people globally do not have access to health care specialists who can diagnose one of the most common childhood illnesses, otitis media (middle ear infection). Affecting more than half a billion people annually, early and accurate diagnosis can ensure appropriate treatment to minimize the widespread impact of ear infections. We developed an image-analysis classification system that can diagnose otitis media with an accuracy comparable to that of general practitioners and pediatricians. The software system, which can be cloud-based for remote image uploading, could provide rapid access to accurate diagnoses in developing countries.