Pardede, Jasman and Utami, Dina Budhi and Rochman, Adlan Chosyiyar (2017) Implementasi "Principal Component Analysis - Scale Invariant Feature Transform" pada Content Based Image Retrieval. Jurnal Teknik Informatika dan Sistem Informasi, 3 (3). ISSN 2443-2229

Preview

Text 12. Implementasi Principal Component Analysis - Scale Invariant Feature Transform Pada Content Based Image Retrieval_Jutisi.pdf Download (1MB) | Preview

Abstract

Content Based Image Retrieval (CBIR) is an image searching technique from a huge image database by analyzing its features. The features can be the color, texture, shape, etc. The method used in this research is a combination of Principal Component Analysis and Scale Invariant Feature Transform ( PCA-SIFT method ). SIFT method is used to detect and describe local features while PCA is used to reduce the dimension of the image. The value of dimension becomes a specific problem in the calculation. The PCA method is applied for the projection of high dimension to low dimension of image. Previously the PCA and only PCA has been frequently applied for digital image retrieval. The searching result is obtained by comparison of the key point descriptor of the query to those of the database. The result of image searching using Wang dataset, indicated that the CBIR using the PCA-SIFT method can reach 90.00% of accuracy and 18.00% of recall.

Actions (login required)

View Item