Prediction of Diabetic Retinopathy from ERG Signal

One of the main causes of visual impairment in diabetics is diabetic retinopathy (DR). Early detection is critical to prevent severe complications, including vision loss. Electroretinography (ERG) signals provide valuable insights into retinal function and are increasingly used for DR diagnosis. In order to evaluate ERG data, extract useful characteristics, and identify anomalies linked to DR, this study makes use of the Discrete Wavelet Transform (DWT). DWT efficiently decomposes ERG signals into time-frequency domains, enabling precise identification of patterns indicative of retinal dysfunction. The proposed system consists of three key stages: signal acquisition, feature extraction using DWT, and classification. Preprocessed ERG signals are subjected to DWT to obtain coefficients, which are analyzed to identify energy variations and abnormalities. A threshold-based or machine-learning classification algorithm is employed to diagnose the presence of DR. The system features a real-time display on an LCD, providing immediate diagnostic results such as "Normal Retina" or "DR Detected." This ensures that healthcare providers and patients can access quick and reliable insights. The goal of this system is to provide clinics and distant healthcare institutions with an affordable, portable, and dependable early DR detection solution. The integration of signal processing, efficient computation, and user-friendly output highlights the innovation and potential impact of this approach in modern ophthalmology.