Identifikasi Rasa Buah Mangga Gedong Gincu Cirebon Berdasarkan Citra Red Green Blue Menggunakan Jaringan Syaraf Tiruan

Main Authors: Yuliana, Upi; Universitas Islam 45, Whidhiasih, Retno Nugroho; Universitas Islam 45, Maimunah, Maimunah; Universitas Islam 45
Format: Article application/pdf eJournal
Bahasa: eng
Terbitan: Fakultas Teknik Program Studi Teknik Komputer Universitas Islam '45' Bekasi , 2017
Online Access: http://www.ejournal-unisma.net/ojs/index.php/piksel/article/view/1625
Daftar Isi:
  • This study aims to identify the image of Gedong Gincu mango based RGB image using Backpropagation ANN. Data used are 90 images of Gedong Gincu Mango. Then the image of Gedong Gincu Mango were split into two groups of data which unkwon to each other, the training data and data testing. Data training is 75 consist of 25 acid class data, 25 medium class data and the remaining is 25 sweet data. Data testing is 15 which consist of 5 acid class data, 5 medium class data and 5 sweet class data. In this study, the parameters used are the learning rate of 0.1, with a number of epoch as many as 2500, the value of MSE (error) is 0.0001 and with variations in the number of hidden layers that will be tested as much as 2, 5, 10, 15, 20, and 25 neurons. With these parameters will be able to identify a class of Gedong Gincu mango with a success rate of 66.6% in 2500 with a variation epoch, neuron number is 2, the value of MSE (error) of 0,269 and at the 35th second. Keyword : Identification,Neural Network, Gedong Gincu Mango