Diabetic Retinopathy Image Recognition Using AI Technique in Rasberry PI
摘要
Regular eye exams are especially important because diabetes can interfere with eye health. Damaged or abnormal new blood vessels can cause vision loss due to high blood sugar, which may damage the eye. Various neural and layered architectures have been applied using CNN algorithms, trained on publicly available image datasets of diabetic retinopathy. Neural networks have been observed to capture lesion color and texture specific to each disease at the time of diagnosis, similar to human decision-making. Experimenting with different features of diabetic retinopathy as inputs to a convolutional neural network is recommended.