Content-Based Image Retrieval using LIRE API for Sensor Network Database

Abstract

Content-Based Image Retrieval (CBIR) is a technique for searching similar images from the large knowledgebase of digital images. Lucene Image Retrieval (LIRE) is a lightweight open source Java library under GPL permit, which is broadly utilized in CBIR. In addition, to provide multiple common and state of the art retrieval techniques LIRE also allows the usage of several frameworks and platforms. As it is extensible in nature it can be used on various process levels like indexing and searching the image based on different feature levels. Due to its adaptive nature developers and researchers can easily modify Lire according to their needs. It is used to extract the feature of image using multiple algorithms like color Layout, CEDD, FCTH. The evaluation is performed sensor network database, which consists of 500 images that has been sensed and aggregated in realistic conditions. In this paper runtime performance for feature extraction is analyzed for each algorithm.