Improving feature vectors for iris recognition through design and implementation of new filter bank and locally compound using of PCA and ICA
2008 1st International Symposium on Applied Sciences in Biomedical and Communication Technologies, ISABEL 2008, Aalborg, Danimarka, 25 - 28 Ekim 2008, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/isabel.2008.4712612
- Basıldığı Şehir: Aalborg
- Basıldığı Ülke: Danimarka
- Anahtar Kelimeler: Biometric identification, False match rate, Feature vector, Filter bank, Principle and independent component analysis
- Atatürk Üniversitesi Adresli: Hayır
Özet
With a growing emphasis on human identification, iris recognition as a biometric identification has recently received increasing attention. Feature vectors are extracted from iris templates and are used for classification purpose. But efficiency of classification operation depends on exclusivity of feature vectors. We have improved features of iris templates by using new filter bank and applying locally of Principle and Independent component analysis on extracted features. Simulation results show improvement of iris recognition by decreasing false match rate in matching level.