Automatic Age Estimation of Face Image using Fusion of Statistical and Texture Features

Automatic age estimation of the face images has different uses such as forensics, customer relationship management and access security control. For this purpose, different features are extracted, processed and selected and the age is estimated using classification algorithms. Extraction and selection of suitable features are a crucial step in this process of age estimation and it is usually so difficult that the estimation accuracy is significantly depends on this stage. Features that can be used to estimate the age are included in both local features in different parts of the face such as wrinkles and global features such as size, shape and appearance. The purpose of this paper is to increase the accuracy of age estimation using fusion of local and global features. Local features are extracted by Haralik and histograms of oriented gradients algorithms while global features are extracted by active appearance model. It is also considered that which combination of local and global features is suitable for increasing age estimation accuracy. Experimental results on some benchmarks show that the fusion of local and global features can improve the accuracy and in addition, fusion of histograms oriented gradients feathers with active appearance model has more accuracy compared to fusion of Haralik with active appearance model features.