Detection of Retinopathy Diseases Using Convolutional Neural Network Based on Discrete Cosine Transform

Abstract

This master thesis proposes a new approach to detecting retinopathy diseases using a convolutional neural network (CNN) based on discrete cosine transform (DCT). Retinopathy is a common eye disease that can cause vision loss if not diagnosed and treated early. The proposed method combines the power of CNN and DCT to improve the accuracy of detection. The input image is transformed into the frequency domain using DCT, which reduces the amount of noise and emphasizes the important features of the image. Then, the transformed image is fed into the CNN for classification. The performance of the proposed method is evaluated using a publicly available dataset of retinal images. The results show that the proposed method outperforms existing methods in terms of accuracy and computational efficiency. The proposed method has the potential to be used in real-world applications for early diagnosis and treatment of retinopathy diseases.