Glasses Detection for Face Recognition Using Bayes Rules

In this paper, we propose a method to detect, extract and to remove the glasses from a face image. The aim is to achieve a face recognition system robust to glasses presence. The extraction is realised using Bayes rules that incorporate the features around each pixel, and the prior knowledge on glasses features that were learnt and stored in a database. Glasses removal is achieved with an adaptive median filter conducted in the points classified as glasses. The experiments indicate the performance of this method is better than that of Deformable Contour, and are satisfying.