Abstract:- Face detection couldalso be a way of detecting faces from pictures, video footages, etc. There arevarious face detection algorithms; one among the widely used algorithm is thatthe Viola Jones algorithm for object detection. The success rate of thisalgorithm for detecting faces was up to 78.4%. during this paper, we present away , which is an improvisation on the prevailing Viola Jones algorithm. we'veimprovised the algorithm to obviously detect the eyes during a face of bothpeople wearing glasses or not. And also other features of face, The detectionof glasses on the face is completed by training a MLP technique. This algorithmprimarily identifies a face with the presence of eyes, which has improved thedetection rate and today our observations have yielded 97% success .Due to thelarge size of realistic galleries, not only the accuracy but also thescalability of a face identification system must be considered. the mostscalability issues are the subsequent . First, the amount of subjects withinthe gallery are often quite large; hence, common search techniques like bruteforce nearest neighbor employed to match probe faces don't scale well.
Keywords:- MLP, EPOCHS,LBP