Content-based Image Retrieval System for Locating Building in Syiah Kuala University using Android Platform

Syiah Kuala University (Unsyiah) regularly involved in various activities and held national events that encompassed lots of people from another city. The lack of facilities to provide information about Unsyiah's building resulted in the difficulty of visitors to quickly and accurately find information about the building. Based on this problem, an Android based application was built by utilizing Content-Based Image Retrieval (CBIR) method, that can use image as the query for the search engine to locate buildings within the Unsyiah's compound. Image of the building will be used as inputs for the query. The method used for developing the image retrieval application is the Scale Invariant Feature Transform (SIFT). The application was built according to the stages of eXtreme Programming (XP) which involves planning, design, coding, and testing. The application is then tested using the Espresso framework and the System Usability Scale (SUS). The result from SUS test is 79.75 which can be categorized as "good" and thus considered as acceptable and user-friendly. The results of the calculation for Recall, Precision, and F-Measure could reach 100% for some building and just 40% for others, depending on the quality of image used for the query. This method is considered as quite good in terms of recognizing object's features because the application could still able to recognize and retrieve the image sought by user despite the poor-quality of image used as the query.