Facial Gender Analysis using Gabor-DWT Feature Extraction Method

Abstract

Facial Gender Analysis has application of specific gender entry detection, human machine interface for digital marketing, real time targeted advertisement and gender demo-graphic analysis. The facial gender can be predicted by classifi-cation of the texture and unique edges pattern. Gabor filter can extract the edge- texture patterns on the face but has problem of high dimensionality with redundancy. For accuracy enhance-ment, the dimension and redundancy is needed to reduce by proposed technique as maxDWT feature optimization method. The proposed model is evaluated on real life challenging dataset of face as illumination variation, POSE, face profile, age variation and obstruction on face as hat, birthmark, moles, speckles, beard, etc. Results shows that proposed technique far better than existing state of art methods of gender prediction.