An Efficient Image Retrieval of Digital Images Using Content Based Image Retrieval by Fuzzy Based Rule Extraction

Abstract

The key idea of image retrieval system is to browse, search and retrieve images from a large distributed database of digital images in an efficient manner. Traditional method of image retrieval utilizes some methods of adding of metadata such as keywords or description to the images so that retrieval can be done using concept of annotation. Manually searching images from the large database is time-consuming and too expensive job. More over semantic gap between the low level features and high level user requirement remains an additional overhead. In order to overcome these issues we are moving towards the “Content Based Image Retrieval”, where CBIR aims at avoiding the use of textual description to retrieves image. Based on the feature similarities like shape, color, and texture & combined features the image retrieval can be done in an efficient manner. Contributing towards the image search from sensor based collected image dataset using CBIR can be helpful for various precautions from Natural disaster. Here we provide the interface between the sensors collected image dataset storage in the Google street view image repository and efficient image retrieval from the Google street view image repository using CBIR.

Keywords: Semantic gap, Annotation, Content based image retrieval, Google street view API