<![CDATA[Gender classification based on facial images is a significant challenge in the field of computer vision, especially when dealing with unstructured data sourced from social media platforms. This study proposes an integrated approach combining facial image preprocessing, clustering methods, and deep learning to enhance the accuracy of gender classification. The dataset used was obtained from a Big Data Competition and consists of male and female face images sourced from Instagram. Preprocessing was performed using OpenCV for face detection and cropping. Subsequently, the data were clustered using K-Means and DBSCAN algorithms to reduce noise and redundancy. Gender classification was then conducted using a sequential learning model based on Inception_v3, enhanced with Agglomerative Clustering for feature refinement. The evaluation of the system demonstrated strong performance with an accuracy of 92.97%, F1-score of 0.89556, precision of 0.97727, and recall of 0.83069. These results confirm that the integration of clustering techniques and deep learning significantly improves the effectiveness of gender classification based on facial images, especially for open-source and non-curated datasets.]]>
Copyrights © 2025
