Gender Prediction based on Facial features Using Convolu-tion Neural Network


  • mais saad safoq university of kerbala


Gender Prediction, Convolutional Neural Network, Facial Features, K-fold cross-validtion


The gender prediction involves the detection of an individual's gender by facial features analysis. The growth of applications that require facial recognition has created an urgent need for such techniques for security and commercial rationales. Gender identification via facial recognition has received great interest among researchers, as well as various techniques used in the field of artificial intelligence and machine learning, with a particular focus on the use of convolutional neural networks (CNNs) for gender classification task. This paper proposes an excellent convolutional neural network (CNN) architecture for gender identification based on facial features. The model is trained and evaluated using a dataset sourced from Kaggle. In addition to the suggested Convolutional Neural Network (CNN) model, the performance of a pretrained MobileNetV2 model and InceptionV3 model is evaluated on the same dataset. The CNN model achieved a commendable accuracy rate of 96.28%, while the MobileNetV2 and InceptionV3 models achieved 95.81% and 97.09%, respectively. The k-fold cross-validation is occupied for the CNN model as a trial for enhancing the accuracy rate to achieve 97.75% accuracy.