Age Prediction System Using Advanced Machine Learning Methodologies

Abstract

Now a days, constructing automatic system for human age estimation is an interesting area for researchers as it covers a wide range of life activities such as authentication and identification, and many other lifetime aspects. Human face is the best part of body that estimates the age of a person based on the shape and appearance of a female or a male. The diverse face characteristics of people and aging difference still provide several unique features which encourage researchers to deal with it. This research proposed a robust system for age estimation based on hybrid machine learning techniques. The human face image is preprocessed firstly to keep the valuable information only within the system and discard the other. Feature extraction will be applied to images using LDA algorithm and the human face images are grouped into three age classes depending on the number of features extracted from each image. Then, the output images will be optimized using hybrid feature optimization technique adopting two inspired algorithms, firefly and Harmony algorithms. The obtained features are classified using J48 classification algorithm. The achieved results showed that the proposed system performed precision of 91.3% and MAE by 1.06.