Jurnal Ilmiah Kursor
Vol 7 No 4 (2014)

IMPRESSION DETERMINATION OF BATIK IMAGE CLOTH BY MULTILABEL ENSEMBLE CLASSIFICATION USING COLOR DIFFERENCE HISTOGRAM FEATURE EXTRACTION

Hani Ramadhan (Informatics Engineering, Faculty of Information Technology, Institut Teknologi Sepuluh Nopember)
Isye Arieshanti (Informatics Engineering, Faculty of Information Technology, Institut Teknologi Sepuluh Nopember)
Anny Yuniarti (Informatics Engineering, Faculty of Information Technology, Institut Teknologi Sepuluh Nopember)
Nanik Suciati (Informatics Engineering, Faculty of Information Technology, Institut Teknologi Sepuluh Nopember)



Article Info

Publish Date
08 Dec 2014

Abstract

IMPRESSION DETERMINATION OF BATIK IMAGE CLOTH BY MULTILABEL ENSEMBLE CLASSIFICATION USING COLOR DIFFERENCE HISTOGRAM FEATURE EXTRACTION aHani Ramadhan, b Isye Arieshanti, cAnny Yuniarti, d Nanik Suciati a,b,c,d Informatics Engineering, Faculty of Information Technology, Institut Teknologi Sepuluh Nopember (ITS) E-Mail: hani.its.042@gmail.com Abstrak Hampir setiap orang akan memperhatikan impresi busana yang dipakai, termasuk busana dengan motif batik. Namun, perpaduan berbagai motif dan warna batik memberikan impresi yang beragam. Sehingga, penentuan impresi dari satu kain batik menjadi sulit. Untuk membantu seseorang dalam menentukan impresi dari busana batik yang dipilih, dibutuhkan sistem yang mampu mengklasifikasikan impresi citra kain batik secara otomatis. Akan tetapi, pembuatan sistem klasifikasi label jamak merupakan memiliki tantangan tersendiri. Penelitian sebelumnya membuktikan bahwa metode klasifikasi ansambel label jamak dengan pencarian threshold mampu menjawab tantangan tersebut dengan kehandalannya dalam menangani himpunan data label jamak. Studi ini bertujuan untuk mengembangkan sistem yang menerapkan metode klasifikasi ansambel label jamak untuk menentukan impresi citra kain batik. Sistem ini memanfaatkan fitur tekstur dan warna yang dihasilkan dari Histogram Perbedaan Warna. Hasil uji coba metode ini memberikan performa yang baik dalam evaluasi label jamak. Nilai evaluasi tersebut antara lain Hamming Loss sebesar 0,173 dan Average Precision 0,866. Kata kunci: Histogram Perbedaan Warna, Impresi Citra Kain Batik, Klasifikasi Label Jamak Abstract Many people will consider the fashion products’ impression that will be worn, including the one with batik motif. Unfortunately, diverse impressions could be produced from combinations of the motif and color from a single batik cloth. Therefore, impression determination becomes a difficult case. To overcome this difficulty, an automatic batik cloth multi-impression classification system should be necessary to aid in choosing certain batik cloth. Nevertheless, this system implementation has its own intriguing challenge. Previous researches implied that multilabel ensemble classification method could deal with the problem against the highly imbalanced dataset. Thus, the aim of this study is to develop the multilabel classification system, which features come from the color and texture feature by Color Difference Histogram. From the test, this method demonstrated good performance by several multilabel evaluations, which are 0.173 by Hamming Loss and 0.866 by Average Precision. Keywords: Color Difference Histogram, Batik Cloth Image Impression, Multi-Label Classification.

Copyrights © 2014






Journal Info

Abbrev

kursor

Publisher

Subject

Computer Science & IT Decision Sciences, Operations Research & Management

Description

Jurnal Ilmiah Kursor is published in January 2005 and has been accreditated by the Directorate General of Higher Education in 2010, 2014, 2019, and until now. Jurnal Ilmiah Kursor seeks to publish original scholarly articles related (but are not limited) to: Computer Science. Computational ...