| Summary: | Face recognition in computer vision has improved in line with the needs of
today's society. However, in its development there are still some kinds of
problems, besides the problem of computing and data storage capacity, the
condition of the human face image into the input system is also an important
issue. Some important aspects that affect the condition of the human face images
include lighting, facial expressions and change attributes such as a mustache,
beard or glasses.
Local Color Histogram (LCHs) is one method that can be used in the
process of face recognition. This method is relatively simple to use in face
recognition process because it is not too much use of the computing process in
comparison with other methods such as using eigenface, PCA and fisherface. In
many LCHs method partitions (blocks) as well as facial image quantization
(acuity level) greatly affect the success rate of color face recognition.
In the process of face recognition using Local Color histograms (LCHs),
the first step face image detected by using YCbCr color segmentation followed by
a facial recognition process by using Local Color histograms (LCHs). Then the
system was tested using the image database of employees of PT. Sarindo
Nusapertiwi (SNP) and ATT with recognition method using a city block
(Manhattan).
The success rate with the introduction of methods LCHs 4x4 block
between 60% -82.86% and the method LCHs 5x5 block between 68.57% -
71.43% using PT SNP data and 86.67% -91.11% by using image of ATT.
|
|---|