Efficient Content Based Image Retrieval of Texture and Natural Images

Abstract

Digitization of various applications has increased very rapidly during last few years This has resulted in the generation of huge amount of multimedia data It is a very crucial task to manage and retrieve any information from this multimedia data with the acceptable accuracy and optimal retrieval time Initially Text Based Image Retrieval TBIR techniques were developed for image retrieval Images are initially retrieved on the basis of textual queries These keywords behave like contents of the images like the file name alternative tags caption of the image etc Natural language processing NLP and concept of bags of words are the two best approaches to text based image retrieval Lycos Multimedia search and Google image search are the most popular image search engines A basic limitation with these search engines is to always have a perfect text query maker This text based query then processed by text engine This engine separates the text into tokens These tokens will be matched with the stored database tokens for the image In most of the cases results of TBIR are not satisfactory due to the semantic gap between the token supplied by the user and tokens of the stored database images Image annotation query formulation language dependency are the major issues with the TBIR based approaches All the images and videos are annotated by some text and this text will be mapped in the user intention at the time of similarity computation Text matching is fast among all approaches to information retrieva. This annotation is laborious if that database size is large and it requires an efficient database annotator newlineIssues with the TBIR are addressed and solved by the Content based image retrieval CBIR approach CBIR came in existence between the years 1994 2000 and can be thought of as the initial phase of research and development on image retrieval by content In this approach a query image is supplied to the retrieval system instead of supplying the text Retrieval system