Abstract:
Digital Libraries mainly concern with learning materials like books, journals, serials, databases, and many others in digital form. In all these cases we are mostly talking about the text materials but never give importance to other forms of information like images, videos, animations, sound, etc. although that constitute a substantial part of the content. This paper focuses on techniques for retrieval of images from digital collection of library being free from their text descriptors. As images are the most important form information that can be attractive in terms of colors, theme, design, and also easy form to remember. Now when searchers are searching for images they do not want the other things to come as search result like text, videos and sounds etc. Past research shows that more sophisticated and effective model for searching images is required so that a searcher will get the perfect result in very less amount of time. The under edge thing is that when a searcher inputs keywords the search engines will only search those images that are annotated with some text descriptors. The authors have done a deep literature survey and studied retrieval mechanisms of various search engines, and the technologies they have used like html, java script, xml, j-query, ajax, j-son, etc. They have tested many technologies and come with a new model for optimized search for images. In this optimized model image is used as input rather than text in the search mode to retrieve the images matching with the input from the digital database developed in library.
