Tez özetleri Astronomi ve Uzay Bilimleri Anabilim Dalı


Computer Aıded Face Recognıtıon System



Yüklə 1,46 Mb.
səhifə244/367
tarix05.01.2022
ölçüsü1,46 Mb.
#64830
1   ...   240   241   242   243   244   245   246   247   ...   367
Computer Aıded Face Recognıtıon System

The process that we generalize as face recognition is a process that is made in two steps indeed. First step, which is defined as face detection is the harder step. If all the areas except face are named background, this step is the process of removing the background and reaching the pure face. The second step is named face recognition or face verification in the literature. The process made here is to determine the face’s owner. Briefly the question that system must answer is “Is there any face, if so where is it?” in face detection step, but “Whose face is this face?” in face recognition step.

In this thesis, a research is made about face recognition problem and the basic methods are investigated. Also, 2 systems that make detection and recognition steps are used. For face recognition step, it is benefited from Neural Networks. This network is educated with 450 non-face figures and 100 face figures. In the figures, for cutting the non-face areas, masking is made. With the testes, the effect of scaling which is a pre-process step is investigated for face recognition. It is determined that, a generalization can’t be made for scaling interval and scaling number so these values must be determined by experiences.

In face recognition “only one sample image per class problem” which is a basic problem of that step is investigated. For this, the success of the method Singular Value Decomposition- SVD that is recommended in the literature is investigated. The pictures which are used in the education set are 112*112 pixel dimension and 8 bit grey level resolution. These pictures were arranged before, by some image processing techniques according to the criteria like ligthing and the angle of face. The general recognition success obtained with this data set is 60 %.


ŞAMLI Rüya ,

Danışman : Doç. Dr. Sabri ARIK

Anabilim Dalı : Bilgisayar Mühendisliği

Mezuniyet Yılı : 2006

Tez Savunma Jürisi : Doç. Dr. Sabri ARIK (Danışman)

Prof. Dr. Ahmet SERTBAŞ

Prof. Dr. Mahmut ÜN

Prof. Dr. Osman Nuri UÇAN

Yrd. Doç. Dr. Oğuzhan ÖZTAŞ




Yüklə 1,46 Mb.

Dostları ilə paylaş:
1   ...   240   241   242   243   244   245   246   247   ...   367




Verilənlər bazası müəlliflik hüququ ilə müdafiə olunur ©muhaz.org 2024
rəhbərliyinə müraciət

gir | qeydiyyatdan keç
    Ana səhifə


yükləyin