Exploring Machine Learning Models for Age Recognition

In recent years, there has been an increasing emphasis on improving artificial intelligence models, particularly those geared at classifying and detecting things, especially human faces. These developments have made a substantial contribution to addressing real-world problems. Numerous studies have been conducted using neural networks, which are commonly referred to as “black boxes” due to their complex decision-making mechanisms and ability to forecast particular characteristics such as age. However, lack of accurate real-world datasets with different variation of images and lack of high-performance processors have caused to predicting accurate age of a person. Consequently, creating accurate and interpretable age recognition models becomes challenging. This journal explores machine learning models that predict a person's age from facial images, combining both classic machine learning methods and neural networks. The research uses a facial image dataset published by the Imperial College of London in the UK. The image characteristics are acquired by the Cannes edges technique, and the dataset dimensions have been reduced using principal component analysis (PCA). Then Random Forest (RF), Support Vector Machine (SVM), Decision Tree (DT), and Convolutional Neural Network (CNN) have been used to determine the most precise model for predicting a person’s age. Each model has been evaluated using performance metrics, confusion matrix, accuracy curve, and loss curve. The random forest algorithm produced the most accurate model with an accuracy of 60%, while the SVM model had the lowest accuracy of 49% compared to other models. CNN without a convolution layer achieved 71% accuracy, whereas CNN with a convolution layer achieved 54% accuracy.