Research And Implementation On Algorithms For Content-based Image Feature Extraction

As the development of Information society, The utilization of images haspermeated all walks of life. The ever growing source of images has provided peoplewith redundant information. The requirement of rapidly searching for useful images isbecoming more and more urgent. The current popular network search engines aremostly based on characters and the research of content-based image search is still inits infancy, The implementation of combing digit image processing, database andimage retrieval technology together to develop fast, convenient image search enginesis of great theoretical and practical values.Content-based image retrieval technology in includes several as aspects: bottomfeature extraction similarity matching and index mechanism. This thesis hascommenced the study on image bottom the classical algorithm of color featureextraction, texture feature, A algorithm of automatic segmentation image is proposed.Based on the comparative and analysis of image feature existed, A new imageretrieval algorithm based on the main color and shape of image blocks was presented.By utilizing feature extraction algorithm mentioned above, great convenience couldbe provided in extracting various features of images, which dramatically increases theimage retrieval accuracy.The present work has used Visual Studio2010as a development tool, and Matlab7.0as experimental tool, to develop a simple content-based image retrieval system.The designed CBIR system includes elements not only the feature extraction of imagewhich is based on the above algorithm, but also a personalized search method thatallows searches based on parameters inputted by customers. A set of experiments havebeen carried out on the system, and some good effects are obtained.