A Novel Method to Predict Age Using Human Eyes in a Deep Learning Approach

Recently, researchers have turned their focus to predict age of people since numerous applications depend on facial recognition approaches. In medical field, Alzheimer disease mainly relies on patients’ ages. Numerous methods have been implemented and developed to predict age. However, these approaches are lack of accuracy because every image has its own unique features such as shape, pose and scale. In Saudi Arabia vision 2030, providing quality of life is one of the twelve initiatives that have been launched.  The health sector has gained attention as the government has placed instructions and rules based on ages to save lives of its elderly residents. Currently, these residents are advised to be vaccinated urgently against Covid-19 based on their ages. In this paper, proposing a practical, consistent, and trustworthy method to predict age is presented. It uses color intensity from face recognition approach and a Convolutional Neural Network to predict age in real-time. Using a segmentation algorithm is involved since the approach takes its input from a video or an image. Several experiments have been conducted on MATLAB to verify and validate its results and their relative errors. A Dataset from the Kaggle website is utilized for different ages as they range from 20s to 50s. This dataset includes over 270,000 of images and its size is roughly 2GB. Consequently, the proposed approach produces ±3.7 years of Mean Square Error (MSE) for the predicted ages. Lastly, a comparison evaluation between literature works and the presented algorithm is provided as well. It shows that the developed approach in this article outperforms in the accuracy since it reaches nearly 90%.