Citation Number
3
Views
35
Downloands
2
Audio Listening
1
Kişilerin davranışlarına, fiziksel özelliklerine bağlı olarak geliştirilen biyometrik sitemler son yıllarda aktif olarak kullanılmaktadır. Kişinin benzersiz özelliklerine dayanan biyometrik sistemler içerisinde yüz tanıma fiziksel temasa gerek duymaması sebebi ile önemli bir yer kaplamaktadır. Bu çalışmada derin öğrenme tabanlı yüz tanıma ve yüz ifadesi tanıma uygulaması gerçekleştirilmiştir. VGG-16, AlexNet ve ZF Net mimarileri ile geliştirilen modeller eğitilerek başarı oranları karşılaştırılmıştır. En başarılı model %92,03 başarı oranı ile VGG-16 mimarisi referans alınarak geliştirilen model olmuştur.
Biometric sitems developed depending on the behavior of individuals, their physical characteristics have been actively used in recent years. In biometric systems based on the unique characteristics of a person, face recognition occurs an important place because it does not require physical contact. In this study, a deep learning-based face recognition and face expression recognition application was carried out. Models developed with the VGG-16, AlexNet and ZF Net architectures have been trained and success rates have been compared. The most successful model was the model developed by reference to the VGG-16 architecture with a 92,03% success rate.
Biometric systems developed depending on the behavior and physical characteristics of individuals have been actively used in recent years. Facial recognition occupies an important place among biometric systems based on the unique characteristics of the person, as it does not require physical contact. In this study, facial recognition and facial expression recognition based on deep learning were implemented. Models developed with VGG-16, AlexNet and ZF Net architectures were trained and their success rates were compared. The most successful model was the model developed based on VGG-16 architecture with a success rate of 92.03%.
Field : Fen Bilimleri ve Matematik; Mühendislik
Journal Type : Uluslararası
Article Detail 
Citation Owners
Similar Articles
Download PDF
Journal Information
Listen Paper
Print This Page
Share
Mendeley
