Human Age Recognition Method Based on Facial Images Using an Ensemble of Neural Network Classifiers

This paper proposes a method for recognizing the age of a person from facial images by using an ensemble of neural network classifiers. The objective of this research is to enhance the effectiveness of human age recognition by employing a neural network classifiers ensemble. The developed method has several advantages: the input image is not required to be square, thereby expanding its applicability; the “convolutional layer – pooling layer” pairs count is calculated empirically, enhancing the neural network classifier accuracy; the count of layers’ planes is calculated automatically, speeding up the determination of the neural network classifier’s structure; the use of an ensemble of neural network classifiers allows for high probability age range classification of individuals. Compared to the scikit-learn package, the proposed method can explore classification using an ensemble of ANN and Hidden Markov Models rather than just an ensemble of neural networks. In addition, this method can explore classification using an ensemble of ANN and Hidden Markov Models rather than only just Hidden Markov Models when compared to the hmmlearn package. Future research prospects include the utilization of the proposed ensemble neural network image recognition method for various intelligent systems aimed at recognizing human characteristics such as gender and emotions.