Gender plays an important role in many areas like health care, E-commerce. In light of the ongoing growth in the computer vision domain, various applications that depend on extracting biometric data, such as facial gender identification for purposes like access control, security, or marketing, are gaining prevalence. A standard gender classifier necessitates a substantial number of training samples to grasp a wide range of discernible features. Convolutional Neural Networks (CNNs) have established their effectiveness in tasks involving image classification, and Xception, an extension of the Inception architecture, has showcased cutting-edge performance across diverse computer vision applications. This research explores the application of the Xception neural network architecture for gender classification using facial images. The dataset used for training and evaluation consists of diverse facial images representing a broad spectrum of ages, ethnicities, and expressions. To improve the model's capability to capture significant features from facial images, preprocessing methods are employed. The devised model incorporates a pre-trained Xception model, integrated into an ensemble framework comprising Support Vector Machines (SVM), Random Forest, and AdaBoost. Our experimental outcomes indicate that training the model on a set of synthetic images yields performance comparable to existing state-of-the-art methods, which utilize authentic images of individuals. The average classification accuracy for each classifier falls within the range of 94% to 95%, mirroring the performance of previously proposed approaches.