KEMIRIPAN POLA PADA CITRA WAJAH MENGGUNAKAN LOCAL BINARY PATTERN --

ABSTRACT
Object image recognition in the use of authentication and verification
systems has been widely used in today's digital era. The image of the object used
as the authentication and verification process does not only use fingerprints, but
also the user's face. Everyone's face has their own differences, but often the
human eye sees similarities or similarities in the faces of several people even
though they are not related by blood, in the sense that strangers who do not know
each other can have facial similarities. This thesis research aims to find the
similarity of patterns in the facial images of several different people. In this study,
the local binary pattern or LBP method works by finding the middle value of a
kernel by comparing the neighboring distance and the number of neighbors in the
image. The stage begins by using a color image of the faces of two or more people
which will be converted into a gray image and then continues with the feature
extraction stage using the LBP method. With the face image from the LBP, then
the histogram equalization is carried out which will produce a plot of the
transformation curve. From the plot of the transformation curve, comparisons are
made on each face image. The number of similarities of the 6 tested facial images
is 5, where there are 2 grouping data from the same x value and adjacent y value.
Keywords: Similarity, face pattern, local binary pattern.