Article Per Year (5 Year)
p-Index From 2019 - 2024
0.23
P-Index
This Author published in this journals
All Journal Jurnal Ilmiah Teknik Elektro Komputer dan Informatika (JITEKI)
Ni Luh Putu Kurniawati
Universitas Pendidikan Ganesha
Computer Science & IT Electrical & Electronics Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Recognition of Balinese Traditional Ornament Carving Images with Convolutional Neural Network and Discrete Wavelet Transform Ni Luh Putu Kurniawati; Made Windu Antara Kesiman; I Made Gede Sunarya
Jurnal Ilmiah Teknik Elektro Komputer dan Informatika Vol 8, No 4 (2022): Desember
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26555/jiteki.v8i3.24360

Abstract

Balinese carvings are less known to the public due to the lack of information about Balinese carvings. Minimum information about Balinese carvings can be overcome by utilizing advances in information technology in the field of image processing, namely the introduction of Balinese carving patterns. In the pattern recognition model of an image, there are several things that can be analyzed, such as the recognition method used, feature extraction, including the model in preprocessing to reduce noise in a Balinese carving image. In this study, the Convolutional Neural Network (CNN) was used to classify Balinese carving images combined with Discrete Wavelet Transform (DWT) in extracting image features. The introduction was made to 25 categories of Balinese carving ornaments. Tests are generated based on the level of accuracy generated in the testing process. Analysis of the results was carried out on the resulting model, namely the analysis of the combination of CNN with DWT and without DWT. Testing the data set with 212 training data and 129 testing data using all DWT channels. Based on the results of the tests that have been carried out, it is found that using the DWT extraction feature produces a higher testing accuracy value, namely 35.66% for 25 classes and 74, 42% for 3 carving classes. Meanwhile, without using DWT, it produces an accuracy value of 32.56% for 25 classes and 66.67% for 3 carving classes. In future research, it is hoped that there will be an improvement in the data set and good shooting with a balanced and adequate number for the 25 carving classes that have been obtained.
Co-Authors I Made Gede Sunarya Made Windu Antara Kesiman