DIABETIK RETINOPATIYANI DIAGNOSTIKA QILISHDA SUN’IY INTELLEKT MODELLARIDAN FOYDALANISH
Keywords:
Keywords: diabetic retinopathy, artificial intelligence, automated diagnosis, fundus images, Ключевые слова: диабетическая ретинопатия, искусственный интеллект, автоматическая диагностика, изображения глазного дна, Kalit so‘zlar: diabetik retinopatiya, sun’iy intellekt, avtomatik diagnostika, ko‘z tubi tasvirlariAbstract
This article provides a detailed examination of the potential use of artificial intelligence (AI) models in the early detection and classification of diabetic retinopathy (DR), one of the most common and serious ophthalmic complications of diabetes mellitus. The study focuses on a comparative analysis between the results produced by an AI model—based on the EfficientNet-B0 convolutional neural network architecture trained on the publicly available Messidor dataset—and clinical diagnoses made by ophthalmologists. The research was conducted on 60 retinal fundus images obtained from 30 diabetic patients during routine ophthalmological examinations. DR classification was performed according to the extended ETDRS (Early Treatment Diabetic Retinopathy Study) scale, allowing for a detailed stage-by-stage distribution. The AI model’s performance was evaluated in terms of accuracy, sensitivity, and specificity. The findings demonstrate a high level of agreement between AI assessments and expert clinical diagnoses, particularly in the early stages of DR. The study underscores the feasibility and potential of incorporating AI into clinical workflows, especially in areas with limited access to specialized care and in the context of large-scale screening programs.
В данной статье подробно рассматривается потенциал применения моделей искусственного интеллекта (ИИ) для раннего выявления и классификации диабетической ретинопатии (ДР) — одного из наиболее распространённых и серьёзных офтальмологических осложнений сахарного диабета.
Annotatsiya (UZ):
Ushbu maqolada sun’iy intellekt (SI) modellarining diabetik retinopatiya (DR) — qandli diabetning eng keng tarqalgan va og‘ir oftalmologik asoratlaridan biri — ni erta tashxislash va tasniflashdagi imkoniyatlari batafsil ko‘rib chiqiladi. Asosiy e’tibor EfficientNet-B0 konvolyutsion neyron tarmog‘iga asoslangan SI modelining Messidor ochiq ma’lumotlar to‘plamida o‘qitilgan holatdagi natijalari va tajribali oftalmologlar tomonidan qo‘yilgan klinik tashxislar o‘rtasidagi taqqosloviy tahlilga qaratilgan [1, 2].
