Age Classification Based on Relative-Features Derived on Circular and Elliptical Neighborhoods using Machine Learning Classifiers

  • Akarapu Swarna, V. Venkata Krishna, Supreethi K. P.

Abstract

Age classification based on human faces plays a significant role in the real world, such as business intelligence, human-computer communication (HCI), and visual observation. The locally-based approaches of age classification attained significant results; however, the researchers mainly considered circular neighborhoods only. A little work is reported based on elliptical patterns, mainly due to the complexity and high dimensional features. The circular, elliptical LBP(CE-LBP) addressed this issue by summing the vertical and horizontal elliptical pixel values with the corresponding circular neighborhood pixel value. The CE-LBP has not derived any relationship between circular and vertical or horizontal elliptical LBPs. This paper computed two frameworks, namely circular and vertical elliptical uniform matrix (CVE-UM) and circular and horizontal elliptical uniform matrix (CHE-UM). These two matrices derive the relative frequencies of isotropic and vertical or horizontal anisotropic patterns. To reduce the dimensionality, uniform patterns are derived. The grey level Co-occurrence matrix features are derived from the proposed frameworks. Machine Learning classifiers are used for classification purposes. The experimental results indicate the efficacy of the proposed method with machine learning classifiers for age classification.

Published
2021-12-02
How to Cite
Akarapu Swarna, V. Venkata Krishna, Supreethi K. P. (2021). Age Classification Based on Relative-Features Derived on Circular and Elliptical Neighborhoods using Machine Learning Classifiers. Design Engineering, 1845 - 1868. Retrieved from http://www.thedesignengineering.com/index.php/DE/article/view/7134
Section
Articles