Generating Mosaic Image Using Color Feature Extraction for Content Based Image Retrieval
DOI:
https://doi.org/10.17977/um068v1i92021p685-691Kata Kunci:
cbir, content based image retrieval, image mosaic, local color histogramAbstrak
This research aims to create an image mosaic from a collection of images by utilizing color feature extraction with CBIR (Content Based Image Retrieval) technique. Some of the things to be considered in CBIR are the selection of the color model, how to represent color features, and metrics for calculating the distance between color features. The system consists of two main parts, preprocessing and image mosaic generating. In the preprocessing stage, image quantization is performed on images in the dataset, then the results are saved to a file that will be used in the image mosaic generating stage. We use two types of color quantization, 27 colors and 64 colors. In the image mosaic generating stage, the input image is divided into many small images (called image tiles), the size of which is adjusted to the size of the dataset. To obtain candidate images, the LCH (Local Color Histogram) values of each image tile are calculated using a certain scheme. Those schemes are 4-block separated, 4-block overlapped, and 5-block. MOSS or DOSS algorithm is used to determine the most similar image for each image tile. Based on the experiments that have been carried out, the 4-block overlapped scheme and 64 color quantization provide the most similar image mosaic.
Penelitian ini berusaha untuk membentuk image mosaic dari kumpulan gambar dengan memanfaatkan ekstraksi fitur warna dengan teknik CBIR (Content Based Image Retrieval). Beberapa hal yang menjadi perhatian dalam CBIR adalah pemilihan model warna, bagaimana merepresentasikan fitur warna, dan metric untuk menghitung jarak antara fitur warna. Sistem terdiri dari dua bagian utama, yaitu tahap preprocessing dan tahap generate image mosaic. Pada tahap preprocessing dilakukan image quantization pada gambar dalam dataset, kemudian hasilnya disimpan ke dalam file yang nantinya digunakan dalam tahap generate image mosaic. Kami menggunakan dua macam color quantization, yaitu 27 warna dan 64 warna. Pada tahap generate image mosaic, input gambar dipotong-potong menjadi banyak gambar kecil (disebut image tile), yang ukurannya disesuaikan dengan ukuran dataset. Untuk mendapatkan kandidat gambar, diambillah nilai LCH (Local Color Histogram) dari setiap image tile dengan menggunakan skema tertentu. Skema pembagian tersebut adalah 4 block separated, 4 block overlapped, dan 5 block. Kemudian ditentukan satu gambar yang paling mirip untuk setiap image tile, dengan menggunakan algoritma MOSS atau DOSS. Dari percobaan yang telah dilakukan, skema 4 block overlapped dan 64 color quantization menghasilkan image mosaic yang paling mirip.
Referensi
Jain, S., Pulaparthi, K., & Fulara, C. (2015). Content based image retrieval. Int. J. Adv. Eng. Glob. Technol, 3, 1251-1258.
Smeulders, A. Worring, M., Santini, S., Gupta, A., & Jain, R. (2000). Content-based image retrieval at the end of the early years. IEEE Transactions on pattern analysis and machine intelligence, 22(12), 1349-1380.
Bheda, D., Joshi, M., & Agrawal, V. (2014). A study on features extraction techniques for image mosaicing. International Journal of Innovative Research in Computer and Communication Engineering, 2(3), 3432-3437.
Markuš, N., Fratarcangeli, M., Pandžić, I. S., & Ahlberg, J. (2015, September). Fast rendering of image mosaics and ascii art. In Computer Graphics Forum (Vol. 34, No. 6, pp. 251-261).
Ghosh, D., & Kaabouch, N. (2016). A survey on image mosaicing techniques. Journal of Visual Communication and Image Representation, 34, 1-11.
Vaghela, D., & Naina, P. (2014). A review of image mosaicing techniques. arXiv preprint arXiv:1405.2539.
Srivastava, D., Wadhvani, R., & Gyanchandani, M. (2015). A review: color feature extraction methods for content-based image retrieval. International Journal of Computational Engineering & Management, 18(3), 9-13.
Liu, G. H., & Yang, J. Y. (2013). Content-based image retrieval using color difference histogram. Pattern recognition, 46(1), 188-198.
Zhang, Y., Nascimento, M. A., & Zaïane, O. R. (2003, July). Building image mosaics: an application of content-based image retrieval. In 2003 International Conference on Multimedia and Expo. ICME'03. Proceedings (Cat. No. 03TH8698) (Vol. 3, pp. III-317). IEEE.
3.png)
1.png)
1.png)
