A Culture-Oriented Image Search System for Indonesian Cultural Paintings with Semantic Multi-Query Analytical Function

The exchange of digital image on internet increases the number of digital image creation in various categories. This condition extends the need of image search system especially for culture-oriented image that treasures set of impressions that makes image searching become more complex. This paper presents a culture-oriented semantic multi-image query system with an analytical function to generate the representative query impressions. This multi query method permits user to attach more than one image queries as their intentions. The set of steps in the analytical function contain several computation methods to extract and generate the dominant query impressions by generating the representative query colors. The representative impressions will consider user intentions to measure the similarity with image database to find the highest degree of similarity of the retrieved images. For experimental study, we implement our system to 248 Indonesian cultural images consisting of five categories of painting styles which are abstractionism, naturalism, expressionism, realism and romance.