Age classification based on appearance model using local ternary direction pattern approach

Abstract

The appearance model play a vital role in many applications related to facial images. This paper derives a new approach of appearance model using local ternary derivative patterns on human facial images for effective age groups classification. In the literature direction patterns are derived with respect to central pixel of the neighborhood. This paper derives ternary direction patters (TDP) between sampling points of the neighborhood with a strong assumption that the relationship between adjacent pixels derive rich information. This paper divides the neighborhood into vertical and horizontal units and derives the TDP and based on the relative frequencies of horizontal and vertical TDP, this paper derives horizontal vertical direction unit matrix (HVDUM). The gray level co-occurrence matrix (GLCM) features are derived on HVDUM for age classification and the experimental results are compared with the existing methods and the results indicate the efficiency of the proposed method over the existing methods.