IMPLEMENTATION OF DCGAN METHOD TO GENERATE BATIK MOTIF

• Main Author: Yoga Adi Prawira, Darien
• Format: Final Project
• Language: Indonesia
• Subjects: Teknik (Rekayasa, enjinering dan kegiatan berkaitan)
• Online Access: https://digilib.itb.ac.id/gdl/view/48081
• Institution: Institut Teknologi Bandung

Batik is one of the respected cultures in Indonesia, especially in Java. The use of traditional technology has made batik difficult to compete with other textile industries. One of the technologies that threatens the traditional batik industry is printing technology because of the easy and inexpensive process used. Therefore, we need a way to shorten the making of traditional batik. The method used in making this model is Deep Convolutional Generative Networks that uses two neural networks of machine learning. This method allows the production of batik motifs based on the learning dataset. The training conducted in making this model uses data collected from the internet with different resolutions and forms of data. The output of this model is batik motifs produced with 6,43 loss value and training time 13 minutes and 51 seconds and based on a survey of 51 people with 8 of the total 64 motifs voted more than 30% voters can be categorized as motifs with flawless results. and resembles the dataset and 10 motifs voted more than 30% voters have defects in coloring but have motifs that resemble the dataset.