Extracting features and content of images to model a data warehouse

Abstract

Data Warehouse is an information system where recent and historic data, related to a topic are unified at a common place. The data is used by companies for their strategic decision making. With the increase in growth of data in the form of images over World Wide Web (WWW), there is a need to organize such data that can help us for decision making. This motivates us to build an image data warehouse. The paper proposes an architecture to model an image warehouse based on the content, metadata and features extracted from images. The metadata as well as images description used as sources of the warehouse are specified by various standards like MPEG-7, JPEG, GIF.