Use of Textural and Structural Facial Features in Generating Efficient Age Classifiers

This paper presents a computational and feasible solution to age classification in humans using facial feature parameters. Computer vision and machine learning have its relevance in research and industry because of its potential applicability and ease of implementation with hidden convolution. Human age classification has gained significant importance in recent times with technological advancement and widespread use of Internet services. In this paper, we propose a methodology that implements chronological growth parameter as well as the textural factor. The facial structural parameter has its significance in the early phase of growth (0–20), whereas texture has its importance in the later span of life (25 onwards), when fine lines start forming in selective facial regions. The proposed method is implemented using statistical techniques (Euclidean distance) and Local Binary Pattern (LBP) for structural and textural feature extraction. As compared to previous work, the parameters used for structural and textural are more in number that contributes to better prediction and classification in these age groups. The results obtained are noteworthy and significant.