Rancang Bangun Mesin Pencari Citra Dengan Pendekatan Temu Balik Berbasis Konten

Abstract

This research aimed to develop the image search system using content-based retrieval approach by applying algorithm to extract low-level feature from images. Low-level information that is used to extract image features in the image retrieval system is color information contained in the image. Composition of color pixels in an image can be represented in a color histogram. The degree of similarity between the image colors is determined based on the distance between the histogram using Laplace distance and Euclidean distance. The smaller the distance between the histogram, the higher the percentage of similarity of an image. Results of image retrieval through the database will be displayed according to the query by example method performed by a user, based on the sequence of images that have a higher percentage of similarity. The test results for images with a maximum size of 150 x 150 pixels get fairly accurate results with percentages of similarity is above 60 percent.