%0 Thesis
%9 Skripsi
%A Suprayitno, NIM.: 06650006
%B FAKULTAS SAINS DAN TEKNOLOGI
%D 2012
%F digilib:54644
%I UIN SUNAN KALIJAGA YOGYAKARTA
%K Pengenalan Daun, Jaringan Syaraf Tiruan, Backpropagation
%P 85
%T SISTEM IDENTIFIKASI DAUN PADA TUMBUHAN OBAT BERBASIS JARINGAN SYARAF TIRUAN
%U https://digilib.uin-suka.ac.id/id/eprint/54644/
%X Human can easily recognize the different types of plants visually. But in  the need of automation, the computer needs to be included in an introduction  system. For this purpose, the system used is Artificial Neural Network. This  research focuses on epoch, and how much is a certain image part can give the  results with an efficient training time and a high success percentage. The more the  epoch amount is, the less time required in the training process.  The method used is to change 100x100 pixel input image to a grayscale  image, and to create binary images. Binary images result will be processed to find  the edge of the leaf pattern from those images, thinning to attenuate the pattern  line, and determining token from the image. Those preprocessing continued with  the training process to develop the system knowledge. A system with enough  knowledge will be given a certain image. A good system is expected to give a  recognized result from the input image.  The system capable to recognize 90% tested images after training the 24  training data that represent 8 types of plants. The best result can be got from the  use of hidden neuron amount as much as 10 neuron, using learning rate of 0.70,  MSE of 0.0243285, neuron input of 16 neuron and using ½ part of input image  that need 9 hours and 54 minutes of training time.
%Z Pembimbing: Shofwatul ‘Uyun, M.Kom