A new integrated approach for CBIR with interactive user feedback using multi features

Abstract

Recent years have seen a swift amplify in the mass of digital image newlinecollections. This ever escalating amount of multimedia data creates a want for newlinenew stylish methods to retrieve the information one is looking for. The predictable newlineapproach alone cannot keep up with the express growth of available data any newlinelonger. But the obstacle is the distinction between the query type, and content type newline(Query as text and content as image) of the searchable objects over the web newlinedatabase. To overcome these search techniques a large amount of search newlinetechniques are proposed and implemented to find the most optimum and efficient newlinesearch for image content in large database. Thus content-based image retrieval newlineengaged many researchers of various fields. There exist many systems for image newlineretrieval meanwhile. newlineFor the purpose of content-based image retrieval (CBIR) an up-to-date newlinecomparison of state-of-the-art low-level feature extraction approach is presented. newlineDue to introduction of web 2.0 users are able to add tags on the images, and in newlinefavor of these texts, most of the users make a text query to find the relevant results newlinefrom the web. This proposed work combines the concept of content based image newlineretrieval and image text to enhance the search results, which promises to provide newlinethe highest accurate search results with respect the traditional CBIR technique and newlineafter implementation of the proposed technique The hybrid approach using text newlinebased search and image content based search keep advantages over traditional newlinesystem considerably. The complete systems are designed and tested for the newlinemedical images. The experimental results demonstrate the performance of the newlineproposed system.