研究目的
To provide a comprehensive review of the anatomy of the human eye, diabetes, diabetic retinopathy, diabetic maculopathy, existing methods for detecting DM in retinal fundus images, performance metrics, and issues in detection.
研究成果
The review concludes that there are various methods for detecting diabetic maculopathy, but some fail in distinguishing exudates and locating fovea and macula in poor quality images. Future work should address these gaps using deep learning and machine learning techniques.
研究不足
The paper is a review and does not conduct original experiments, so it has limitations in terms of providing new data or experimental validation. It relies on previously published works, which may have their own limitations such as dataset variability and algorithm performance issues.
