<![CDATA[This study aims to develop a gender classification model based on eye images using the Support Vector Machine (SVM) algorithm. The dataset consists of 13,499 eye images divided into two classes: male and female. The methodology includes preprocessing by converting images to grayscale and resizing them to 64×64 pixels, followed by feature extraction using raw pixel representation resulting in a 4,096-dimensional vector. The data is split into 80% for training and 20% for testing, and SVM parameters are optimized using grid search with 5-fold cross-validation. The SVM model employs an RBF kernel with parameters C=10 and gamma='scale'.Evaluation is carried out using accuracy, precision, recall, F1-score metrics, and a confusion matrix. A decision boundary is visualized using PCA to analyze data separability. The results show excellent performance with 99.96% accuracy, 100.00% precision, 99.95% recall, and 99.98% F1-score. The confusion matrix indicates near-perfect classification, with 648 male samples and 2,051 female samples correctly classified without misprediction. This study demonstrates that the SVM algorithm, even with simple preprocessing, can achieve high accuracy in gender classification based on eye images, showing strong potential for practical implementation in biometric systems]]>
Copyrights © 2026
