Universitas Stikubank (Unisbank) Semarang Repository

DEEP LEARNING OBJECT DETECTION PADA MOTIF BATIK MENGGUNAKAN TENSORFLOW

Pratama, Reza Wisnu (2021) DEEP LEARNING OBJECT DETECTION PADA MOTIF BATIK MENGGUNAKAN TENSORFLOW. Undergraduate thesis, Universitas Stikubank.

[thumbnail of HLM JUDUL] PDF (HLM JUDUL)
Download (795kB)
[thumbnail of ABSTRAK] PDF (ABSTRAK)
Download (120kB)
[thumbnail of BAB I] PDF (BAB I)
Download (188kB)
[thumbnail of BAB II] PDF (BAB II)
Restricted to Repository staff only

Download (1MB)
[thumbnail of BAB III] PDF (BAB III)
Restricted to Repository staff only

Download (1MB)
[thumbnail of BAB IV] PDF (BAB IV)
Restricted to Repository staff only

Download (1MB)
[thumbnail of BAB V] PDF (BAB V)
Restricted to Repository staff only

Download (110kB)
[thumbnail of DAFTAR PUSTAKA] PDF (DAFTAR PUSTAKA)
Download (183kB)
[thumbnail of LAMPIRAN] PDF (LAMPIRAN)
Restricted to Repository staff only

Download (1MB)

Abstract

Batik menjadi warisan kebudayaan Indonesia yang harus dilestarikan. Motif batik Indonesia mempunyai ragam atau karakteristif motif yang berbeda di setiap daerah Oleh karena itu diperlukan suatu system yang otomatis mendeteksi keanekaragaman motif batik. Keanekaragaman tersebut dideteksi atau diklasifikasikan menggunakan Deep Learning dengan metode CNN (Convolutional Neural Network). Tensorflow menyediakan berbagai modul dan library yang mendukung komputasi untuk proses deteksi atau klasifikasi motif batik. Pada penelitian ini menggunakan EfficientNet sebagai base model CNN yang disediakan oleh Tensorflow dengan jumlah citra sebagai dataset ada 750 citra dari 5 kelas motif batik yaitu Motif Batik Betawi, Cendrawasih, Kawung, Megamendung dan Parang. Setelah proses training model dengan 100 epoch selesai dan dilakukan proses prediksi menghasilkan rata-rata akurasi sebesar 86%. Batik is an Indonesian cultural heritage that must be preserved. Indonesian batik motifs have a variety or characteristics of different motifs in each region. Therefore, we need a system that automatically detects the diversity of batik motifs. The diversity is detected or classified using Deep Learning with the CNN (Convolutional Neural Network) method. Tensorflow provides various modules and libraries that support computation for the detection or classification of batik motifs. In this study, using EfficientNet as the base model for CNN provided by Tensorflow with the number of images as a dataset, there are 750 images from 5 classes of batik motifs, namely Batik Betawi, Cendrawasih, Kawung, Megamendung and Parang motifs. After the model training process with 100 epochs is completed and the prediction process is carried out, it produces an average accuracy of 86%.

Item Type: Thesis (Undergraduate)
Additional Information: SKR.I.05.01.2055 NIM 17.01.53.0022
Uncontrolled Keywords: Batik, Deteksi Citra, Deep Learning, Convolutional Neural Network, Tensorflow Batik, Image Detection, Deep Learning, Convolutional Neural Network, Tensorflow
Subjects: Q Science > QA Mathematics > QA76 Computer software
Faculty / Institution: Fakultas Teknologi Informasi > Program Studi Teknik Informatika
Depositing User: Teteh Hayati
Date Deposited: 11 Nov 2021 03:00
Last Modified: 11 Nov 2021 03:00
URI: https://eprints.unisbank.ac.id/id/eprint/8053

Actions (login required)

View Item View Item