A New Method For Face Recognition Using Feature Clustering With Fuzzy Parameters

Abstract

In this paper, we have applied Gabor filter for fiducial point localisation. The fiducial points are represented as trapezoidal fuzzy numbers. These fiducial points are then transformed into crisp numbers. The number of fiducial points are then reduced by using a distance formula. The distance of each of these fiducial points are then stored in the database of the system. The same methodology is applied on the input face which is to be matched with the faces available in the database. Then a fuzzy preference relation matrix is obtained .The largest eigen value of this matrix is then determined. Once the largest eigen value is determined the corresponding priority vector can easily be obtained, from which we can easily match the input face with the database. In the broadest sense we have observed we have used the fuzzy mathematics to counter the impreciseness in facial recognition.

Application of Computer Systems in Facial Recognition Efficiency

“…The two images, however, will definitely not be exactly the same due to different factors [22] such as lighting and camera position, as discussed earlier. To counter these difficulties, logical scoring of preference is used with fuzzy numbers [23]. These are used to be able to classify whether the images are matches or not despite the inexactness of the comparison [24].…”