Publication

Abstract : Recognizing age, gender, and ethnicity is a critical issue in computer vision with several real-world applications, such as social analytics, surveillance and customized user experiences. This paper suggests a method based on deep learning for reliable multimodal gender, ethnicity and age recognition from facial information. This framework focuses on convolutional neural networks (VGG16) to extract discriminative features at different scales, while handling temporal dependencies. This study introduces novel attention mechanisms to highlight salient facial regions for improved recognition performance. Our approach has been extensively tested on benchmark datasets, showcasing its remarkable performance by achieving cutting-edge results in accuracy, robustness, and computational efficiency. It provides 82% accuracy in VGG16 for Age recognition, 95.31% accuracy in VGG16 for Gender recognition, 98.44% accuracy in VGG16 for Race recognition.