Content Based Image Retrieval Busana Muslimah Menggunakan Fitur Kombinasi PHOG dan DCD

Cut Mutia; Muhammad Akmal

doi:10.36294/jurti.v5i1.2021

Content Based Image Retrieval Busana Muslimah Menggunakan Fitur Kombinasi PHOG dan DCD

Cut Mutia, Muhammad Akmal

Abstract

Abstract - The development of Muslim clothing has increased significantly, especially in Indonesia, where the majority of the population is Muslim. Designers have appeared with a variety of fashion styles that attract consumers, especially young consumers. Technological developments have also influenced the way designers promote and sell their work by utilizing online commerce. So that consumers can search and find various models and colors of clothing that interest them. In this study, the search for Muslim clothing was carried out using the visual content of images or also known as features. The features used in this study are a combination of color and shape features, because shape and color are an attraction for consumers in choosing clothes. Search using image content as a query is known as CBIR. This study applies the PHOG as a shape feature extraction method and the DCD for the extraction of color features. The dataset used is 180 images of Muslim clothing. The performance of CBIR Muslim clothing using a combination of PHOG and DCD features is measured with recall, precision, and f-measure parameters. The results of the performance measurement show that the combination of PHOG and DCD features is better applied to blouse pants clothing. The highest scores were obtained from the red clothing group of 68% (P), 22.2%, and 33.3% (F). Then followed by a group of robes, and outerwear.

Keywords - CBIR, PHOG, DCD, Feature and Clothing.

Abstrak - Busana muslimah memiliki bentuk dan warna yang beragam dan menarik minat konsumen, banyak situs belanja online yang mempromosikan dan menjual produknya. Namun, pencarian busana muslimah pada situs tersebut masih menggunakan teks, sehingga hasil pencarian produknya sering tidak sesuai dengan harapan konsumen.Pada penelitian ini, pencarian busana muslimah dilakukan dengan menggunakan isi visual citra atau disebut juga sebagai fitur, adapun fitur yang digunakan adalah kombinasi fitur warna dan bentuk, karena bentuk (model) dan warna merupakakan daya tarik bagi konsumen dalam memilih busana. Pencarian menggunakan isi gambar sebagai query dikenal dengan Content Based Image Retrieval (CBIR). Penelitian ini menerapkan Pyramid Histogram of Oriented Gradient (PHOG) sebagai metode ekstraksi fitur bentuk dan Dominant Color Descriptor (DCD) untuk ekstraksi fitur warna. Dataset yang digunakan ada 180 citra busana muslimah, terdiri dari 60 citra blus celana, 60 gamis, dan 60 outerwear. Kemiripan query dengan dataset dilakukan dengan menghitung jarak Euclidean. Pengukuran Kinerja CBIR busana muslimah menggunakan kombinasi fitur PHOG dan DCD dilakukan dengan menghitung nilai recall, precision, dan f-measure. Hasil pengukuran kinerja menunjukkan bahwa kombinasi fitur PHOG dan DCD lebih baik diterapkan pada busana blus-celana. Berdasarkan pengukuruan kinerja dengan precision (P), recall (R0, dan f-measure (F). Nilai tertinggi diperoleh dari kelompok busana berwarna merah sebesar 68 % (P), 22.2%, dan 33.3 % (F). Kemudian diikuti dengan kelompok busana gamis, dan outerwear.

Kata Kunci - CBIR, PHOG, DCD, Fitur dan Busana.

Keywords

CBIR, PHOG, DCD, Feature and Clothing

Full Text:

PDF

References

Y. Dhyanti, K. Munadi, and F. Arnia, “Penerapan Deskriptor Warna Dominan untuk Temu Kembali Citra Busana pada Peranti Bergerak,” J. Rekayasa Elektr., vol. 12, no. 3, p. 104, 2016.

A. F. Dewi, F. Arnia, and R. Muharar, “Effectiveness of MPEG-7 color features in clothing retrieval,” Bull. Electr. Eng. Informatics, vol. 6, no. 2, pp. 166–173, 2017.

H. Maghfirah, F. Arnia, and K. Munadi, “Temu Kembali Citra Busana Muslimah Berdasarkan Bentuk Menggunakan Curvature Scale Space (CSS),” J. Nas. Tek. Elektro dan Teknol. Inf., vol. 6, no. 1, pp. 74–83, 2017.

C. Mutia, F. Arnia, and R. Muharar, “Improving the performance of CBIR on islamic women apparels using normalized PHOG,” Bull. Electr. Eng. Informatics, vol. 6, no. 3, pp. 271–280, 2017.

W. Zhou et al., “Fashion recommendations through cross-media information retrieval,” J. Vis. Commun. Image Represent., vol. 61, pp. 112–120, 2019.

O. N. Shpakov and G. V. Bogomolov, “Technogenic activity of man and local sources of environmental pollution,” Stud. Environ. Sci., vol. 17, no. C, pp. 329–332, 1981.

A.W. Rosyadi, R. Danardono, S.S. Manek and A. Z. Arifin, “A flexible sub-block in region based image retrieval based on transition region, vol. 11, no. 1, pp. 42–51, 2018.

Y. Park, K. Park, and G. Kim, “Content-based image retrieval using colour and shape features,” Int. J. Comput. Appl. Technol., vol. 48, no. 2, pp. 155–161, 2013.

A. Bosch, A. Zisserman, and X. Munoz, “Representing shape with a spatial pyramid kernel,” Proc. 6th ACM Int. Conf. Image Video Retrieval, CIVR 2007, pp. 401–408, 2007.

H. Shao, Y. Wu, W. Cui, and J. Zhang, “Image retrieval based on MPEG-7 dominant color descriptor,” Proc. 9th Int. Conf. Young Comput. Sci. ICYCS 2008, pp. 753–757, 2008.

N. C. Yang, W. H. Chang, C. M. Kuo, and T. H. Li, “A fast MPEG-7 dominant color extraction with new similarity measure for image retrieval,” J. Vis. Commun. Image Represent., vol. 19, no. 2, pp. 92–105, 2008.

N. V. Shirahatti and K. Barnard, “Evaluating image retrieval,” Proc. - 2005 IEEE Comput. Soc. Conf. Comput. Vis. Pattern Recognition, CVPR 2005, vol. I, pp. 955–961, 2005.

DOI: https://doi.org/10.36294/jurti.v5i1.2021