Jaringan neural Untuk Pengenalan Pola yang Invarian Terhadap Translasi Dan Rotan
| Main Author: | Perpustakaan UGM, i-lib |
|---|---|
| Format: | Article NonPeerReviewed |
| Terbitan: |
[Yogyakarta] : Universitas Gadjah Mada
, 1996
|
| Subjects: | |
| Online Access: |
https://repository.ugm.ac.id/25463/ http://i-lib.ugm.ac.id/jurnal/download.php?dataId=8458 |
Daftar Isi:
- Analog coupled neurons (CN) can be used in trainable pattern recognition systems. A training algorithm called Coupled Neuron Rule (CNR) is used to implement classification functions. A character recognition.system involving an invariance network and a trainable classifier is proposed. The invariance net can be trained to produce a set of outputs that are insensitive to translation and rotation of the retinal input patterns. The outputs of the invariance net are scrambled. When these outputs are fed to a trainable classifier, the final outputs are descrambled and the original patterns are reproduced in standard position and orientation