The automated analysis of ophthalmological images has become a key tool for the early diagnosis of retinal diseases such as diabetic retinopathy and hypertensive retinopathy. This study presents a comparative approach to the segmentation and classification of retinopathies using multiple image processing techniques and machine learning algorithms. Different models based on convolutional neural networks (CNN) were implemented, combining original grayscale images, vascular segmentations, optic disc detection, and texture patterns (LBP). Each model was independently trained and evaluated using metrics such as accuracy, sensitivity, specificity, and area under the ROC curve (AUC-ROC). The results show that the combined use of segmentation techniques and textural features significantly improves the accuracy in retinopathy classification. Additionally, the potential of models based on transfer learning for retinal image analysis is highlighted. These findings provide guidance for the development of more robust and accurate computational systems in the automatic diagnosis of ophthalmological pathologies. This study achieved a classification accuracy of 95.1%, with an F1-score of 94.3%, sensitivity of 92.8%, and specificity of 93.6% using deep learning models combined with preprocessing and feature extraction techniques. These results highlight the potential of the proposed methodology to support automated diagnosis of retinopathies.