Content Based Image Retrieval Using Enhanced Vocabulary

Content Based Image Retrieval is a technique that makes use of the visual content of an picture, to search for similar pictures in large-scale image databases, as per a user’s interest. The CBIR is driven by the requirement to go looking the exponentially increasing set of image and video databases expeditiously and effectively. The visual content of a picture is analyzed in terms of lowlevel options extracted from the image. In CBIR system, it's usual to cluster the image options in three main classes :color, texture and form. Ideally, these options ought to be integrated to supply higher discrimination within the comparison method. For example by constructing an object vocabulary containing query-associative objects by mining frequent object patches from the resulting image collection of the expanded query set. After representing each image as a bag of objects, the retrieval model can be derived from a riskminimization framework for language modelling. In this paper we proposed a technique which uses enhanced vocabulary to map the features extracted from the images and give the relevant result The experimental results show that the proposed methods can significantly outperform the existing approaches.