Handheld retinal camera for AI-based diabetic retinopathy screening

项目总结 美国疾病控制与预防中心建议对美国3420万糖尿病患者进行筛查 每年用于糖尿病视网膜病变(DR)，这是可预防失明的主要原因。不到50%的糖尿病患者 由于缺乏医疗护理和眼科专家，患者实际上遵守了这些指南，时间和 资金紧张，以及缺乏早期疾病的症状。 为了解决这些问题，我们在AI Optics的团队正在开发世界上第一个基于人工智能的 手持视网膜相机，允许进行护理点DR筛查。这款设备设计成便于携带、使用方便 易于使用，且工作流程友好。它对非扩张型患者进行高精度的DR筛查， 实现最佳安全性，并保持对连接问题的弹性。我们的目标是，这种新的诊断设备将 将DR筛查从眼科专家的办公室扩展到初级保健、验光办公室、糖尿病 诊所和零售保健机构。这种更多的筛查将增加早期诊断率 并避免可预防的视力损失。 在这个第一阶段的SBIR项目中，我们将开发一款符合国际标准化组织10940：2009年标准的视网膜相机， 它将能够在手持设备中捕获高质量的眼底图像。以确保全尺寸图像 可以进行分类，我们也将利用我们专有的、深度学习的人工智能系统。 最后，我们将确保从我们的视网膜相机捕获的图像可以通过我们的人工 这一项目的成功完成将带来一个更好的 更易获得的DR筛查工具，可实现更早的DR诊断、防盲和 为数百万糖尿病患者节省了大量成本。

PROJECT SUMMARY The CDC recommends that each of the 34.2 million patients with diabetes in the United States is screened annually for diabetic retinopathy (DR), a major cause of preventable blindness. Less than 50% of diabetes patients actually follow these guidelines due to lack of access to medical care and eye specialists, time and money constraints, and lack of symptoms with early-stage disease. To address these problems, our team at AI Optics is developing the world’s first artificial intelligence-based handheld retinal camera to allow for point-of-care DR screening. This device is designed to be portable, easy to use, and workflow friendly. It performs high-accuracy DR screenings on non-dilated patients, maintaining optimal security and remaining resilient to connectivity issues. Our goal is that this novel diagnostic device will expand DR screenings beyond the offices of eye specialists and into primary care, optometry offices, diabetes clinics, and retail health settings. This increased access to screening will increase early-stage diagnosis rates and avoid preventable vision loss. In this Phase I SBIR project, we will develop a retinal camera that complies with ISO 10940:2009 standards, which will be able to capture high-quality fundus images in a handheld device. To ensure that full-scale image classification can be conducted, we will also utilize our proprietary, deep-learning artificial intelligence system. Finally, we will ensure that images captured from our retinal camera can be analyzed by our artificial intelligence for the presence of DR. The successful completion of this project will result in an improved and more accessible tool for DR screenings that could lead to earlier DR diagnosis, blindness prevention, and significant cost savings for millions of people with diabetes.