Hash Based Secure and Efficient Scaling Approach for Detecting the Diabetic Retinopathy Disorder

Diabetic Retinopathy (DR) is a serious concern to diabetics, resulting in eye issues if left untreated. Conventional DR diagnostic procedures often lack robust privacy protection safeguards, leaving patient data vulnerable to unauthorized access and change. To address this pressing issue, we offer a novel deep-learning model that prioritizes both accurate diagnosis and privacy protection. The strength of the Efficient NetB2 model is leveraged in our proposed model, which incorporates numerous conv2d layers and activation functions to enable precise prediction and categorization of DR stages using retina images. The Secure Hash Algorithm (SHA-512) is used to produce unique hash values for each image, which serve as secure representations of the data. These hash values not only preserve patient privacy but also allow for the rapid identification of any unauthorized alterations to the stored images. This assures the accuracy of the diagnoses, allowing healthcare practitioners to make educated decisions while maintaining patient anonymity. Validate the effectiveness and efficiency of our suggested approach by rigorous experimentation on a large dataset. Integrating accurate DR detection with content change detection based on hash values creates a robust privacy-preserving approach. The proposed model not only mitigates the risk of vision problems for diabetic patients but also guarantees the privacy and integrity of their medical data.