A Novel Method for Facial Features Extraction Using Bilateral Filtering and Wavelet Transform

In this paper we present a new approach to the problem of facial features extraction from a still facial image. The method does not need to process features such as, eyes, nose and mouth separately in different stages of processing. It detects the eyes, nose and mouth region in a face image at a single processing stage, reducing computational cost. Our four step method consists of Bilateral filtering, Wavelet Transform (WT), Dilation, and binary threshold. The proposed method shows high accuracy even in the presence of various gestures, illumination variations, closed eyes, and eyes with glasses.