Research And Implementation Of Feature Based Image Retrieval Technology

Feature based image retrieval is a hotspot of research in multimedia information processing field, and has capacious developed foreground. This technology extracts visual features as the retrieval features, such as color, texture, shape etc. Generally, an example image is needed when a user uses the retrieval system, which can extract the features of the example image, then compare with the features of other images in the database and show the result to the user. Although many feature-extracting methods and similarity measurements have been raised, they are not mature enough to get good retrieval result. New technology of feature based image retrieval researching is required to improve performance.This dissertation briefly summarizes feature based image retrieval technology, and researches some key techniques of the image retrieval which specially focuses on feature extraction. for the shortcoming of the traditional histogram could not express the color spatial distribution, this dissertation propose an improvement color region area histogram retrieval method; for the insufficiency of the edge direction histogram method expression shape information, this dissertation propose an edge direction-corner based image retrieval method. Compare with the traditional methods, the new methods have better retrieval performance, this have proved by the experiment.At present the image retrieval no longer relies on the sole visual characteristic of image, because these basic visual characteristic has only used the image partial information, Therefore the retrieval result is often insufficient the ideal. In order to solve this problem, this dissertation proposed that one kind based on the image characteristic combination's searching algorithm, first normalization of image feature vectors and similarity distance on some characteristic before calculating the similarity distance between two images, then accumulates the weighted similarity distance on various characteristics, by this achievement two image similar distances. It may adapt user's various requirement by adjust corresponding weight. The empirical datum indicated, this method strengthened the retrieval flexibility and raised the retrieval efficiency. Finally, this dissertation has designed an image retrieval experiment system as the goal of take algorithm test, and has given the system detailed design.