Improved Markovian Model for Annotation Based Image Retrieval using Multiple and Synonymous Keywords?

With the rapid growth of digital media, efficient and effective image retrieval techniques are desired. The process of image retrieval fails due to sensory gap. Annotation based image retrieval is one of the promising image retrieval technique because of its power to represent user queries and semantic contents of images. There are many methods presented for automatic annotation, indexing as well as annotation-based retrieval of images. Recently markov chain based method, Markovian Semantic Indexing is introduced which is efficient as compared to all previous methods. In this paper we propose a unique methodology for classification and annotation-based retrieval of pictures. The new methodology is also applicable within the context of an online image retrieval system. The new methodology is shown to possess bound theoretical blessings and additionally to attain higher precision versus Recall results compared to existing technique of MSI in Annotation-Based Image Retrieval (ABIR). The proposed system extends MSI by considering multiple keywords and synonyms in the user query. The experimental results demonstrate the performance of the proposed system.