No Detail Found!

Enhancing a reliable traditional clothes pattern retrieval: cnn model and distance metrics

Author(s)
ABDUL HARIS RANGKUTI Informatics Department, School of Computer Science, Bina Nusantara University, Jakarta-11480 Indonesia.
VARYL HASBI ATHALA Informatics Department, School of Computer Science, Bina Nusantara University, Jakarta-11480 Indonesia.
EVAWATY TANUAR Informatics Department, School of Computer Science, Bina Nusantara University, Jakarta-11480 Indonesia.
JOHAN MULIADI KERTA Informatics Department, School of Computer Science, Bina Nusantara University, Jakarta-11480 Indonesia.
Abstract
This study describes the traditional clothing patterns retrieval very diverse forms and textures. With so many variations in each clothes pattern, an appropriate and supported CNN model is needed as well as the right distance matrix method to support retrieval performance. This research was conducted on 76 types of traditional cloth patterns originating from 22 regions in Indonesia. In conducting an experiment to retrieve this clothes pattern, 3 CNN models were used, namely EfficientnetB7, InceptionV3 and VGG 19 and 2 distance matrix methods, namely Euclidean and Manhattan. Based on the experiment, the average with 2 measurement distances is 86.2%. The CNN model has the highest accuracy on EfficientNetB7 with an average of 92%. Inception V3 with an average accuracy of 85%, and the next VGG 19 has an average of 81%. Basically, some batik cloth patterns have an accuracy of up to 100% retrieval using 3 CNN models, but some patterns have an accuracy below 50%, so this is part of the continuation of research on traditional cloth patterns should convey the importance of your research in a concise and logical manner.
Publication Details
Page(s) 3183-3193
DOI DOI not available
Published Journal: Journal of Theoretical and Applied Information Technology, Volume: 100, Issue: 10, Year: 2022

Request the full text from PASTIC

Keywords
CNN, Inception V3, VGG 19, Euclidean, EfficientnetB7, Manhattan
0

Citations

0

Downloads

80

Views