Novel architecture for CBIR SAAS on Azure cloud

Multimedia messages like images or videos are used in large scale for effective way of communication in today time. Content Based Image Retrieval (CBIR) is an efficient and effective way for retrieval of relevant/similar images from large database based on features extraction principal. In this paper we have proposed a system which extracts images, already uploaded in database through worker role, after analyzing property or message associated with images. The main purpose of this paper is to purpose a system which facilitates quick retrieval of data (images) from cloud. It can be used easily for retrieving images, based on the query image, from a large s et of distinct images. We are proposing one architecture in which feature extraction technique is used to extract image by feature vectors based on color, shape, texture features etc. Mentioned features retrieved from the images can be stored in vectors called feature vectors of images and therefore these are compared to the feature vectors of query image and the image information is sorted in decreasing or ascending order. The Proposed system is an application built on Windows Azure platform for the uploading and retrieval of images.