APLIKASI PENGENALAN SIDIK JARI DENGAN WAVELET SYMLET DAN WAVELET DAUBECHIES MENGGUNAKAN JARINGAN SARAF TIRUAN PERAMBATAN BALIK
| Main Authors: | Tarigan, Ruhi Agatha, Hidayatno, Achmad , Zahra , Ajub Ajulian |
|---|---|
| Format: | Thesis NonPeerReviewed application/pdf |
| Terbitan: |
, 2012
|
| Subjects: | |
| Online Access: |
http://eprints.undip.ac.id/35583/1/L2F006079_MTA.pdf http://eprints.undip.ac.id/35583/ |
Daftar Isi:
- Intdentification of a personal identity by using a password or card is not classified as safety system because the security system is impenetrable when the password or the card is used by other users. Given these shortages, biometric recognition techniques must be developed. Biometric recognition techniques based on the characteristics of human nature, such as the behavioral characteristics and physiological characteristics such as face, fingerprint, voice, palm, iris, DNA, and signatures. The design of fingerprint recognition application uses two-dimensional discrete wavelet transform to feature extraction and artificial neural network majoring back propagation as learning algorithm and the identification of the system. The purpose of this thesis is to analyze the type of Mother Wavelet with the same level that gives the best recognition rate in the system of fingerprint recognition applications. This final thesis use two types of Mother Wavelet, daubechies4 and symlet2 with the same decomposition level, ie level 2. The results from the decomposition of a fingerprint using a wavelet transform will produce coefficient approximation. Tthe greatest Magnitude approximation of coefficients approximation will be used as input for artificial neural network learning. Based on training data results, Mother Wavelet composition that gives the best recognition rate is a daubechies4 (db4) with decomposition of 2, equal to 92%. While in testing the data test, The best recognition rate is a network that uses daubechies4 (db4) with the level of decomposition of 2 is equal to 90.4%. Key words : discrete wavelet transform 2D, fingerprint recognition, back propagation neural network.