Penerapan Metode Convolutional Neural Network untuk Klasifikasi Motif Tenun Ikat Manggarai, Nage-Keo, dan Ngada
Abstract
This study aims to develop a classification application of Flores woven from the Manggarai, Nage-Keo or Ngada areas using the Convolutional Neural Network (CNN) method. The software development process uses an incremental method. The data used in this study are images of Flores woven from the Manggarai, Nage-keo, Ngada, Ende and Maumere areas. Ende and Maumere woven images are used as comparative data to detect images that do not include Manggarai, Nage-Keo, and Ngada ikat weaving. This data will be extracted using the VGG16 model that has been trained with weights in the imagenet data to obtain characteristic values in the form of a multidimensional matrix that can recognize colors, textures, etc. The results obtained in this study, the CNN algorithm can be applied to the classification of ikat motifs with a training precision of 99.87% and a validation precision of 98.66%.