PENGENALAN WAJAH MENGGUNAKAN IMAGE PROCESSING DAN JARINGAN SYARAF TIRUAN SELF ORGANIZING MAPS (SOM)
| Main Author: | BUDIHARJO, AGUNG |
|---|---|
| Format: | Thesis NonPeerReviewed Book |
| Bahasa: | ind |
| Terbitan: |
, 2014
|
| Subjects: | |
| Online Access: |
https://eprints.untirta.ac.id/12471/1/PENGENALAN%20WAJAH%20MENGGUNAKAN%20IMAGE%20PROCESSING%20DAN%20JARINGAN%20SYARAF%20TIRUAN%20SELF%20ORGANIZING%20MAPS%20%28SO.pdf https://eprints.untirta.ac.id/12471/ https://ft.untirta.ac.id/id/ |
Daftar Isi:
- Face recognition technique using artificial intelligence in biometrics become a challenging issue in the last few years. Face recognition is performed using 6 image basic human face configuration such as. normal, happy, sad, angry, smile, and surprised. This technique is expected to replace conventional identification techniques were judged not quite reliable in its utilization. This face identification system using artificial intelligence approach to mimic base face redundancy through reduction of facial image data is accomplished through the use of image compression stage 2-dimensional discrete cosine transform (DCT). The DCT extracts features from face images based on skin color. Feature vectors are constructed by computing DCT coefficients. Self-organizing map (SOM) using unsupervised learning method for classifying DCT-based feature vectors into several groups and identify if the subject on the face image data input was “present” or “not present” exists in the image database. Identification is performed trough the database using MATLAB image as much as 60 face image, consist of 10 people with each 6 mimic face image. The main purpose of this identification is to determine the precise parameters in developing a system of identification using artificial neural network SOM to the face recognition. Recognition rate highest obtained on epoch 200 and 500 equal to 100 %. Testing is by changing the parameters of alpha 0.01, 0.1, and 0.5. The highest accuracy on a parameter epoch 500 and learning rate 0.5 of 91%. So that it can be concluded that the face recognition with a method of artificial neural network SOM can be conducted to identify the face based on mimic face