| Summary: | 碩士 === 朝陽科技大學 === 資訊管理系碩士班 === 91 === Recently, how to efficiently manage image data in multimedia databases has gotten many attentions. Image similarity search is employed to retrieve similar images from a database. Colours are one of the visual features that people immediately perceive when looking at an image. Colour histograms, which encompass colour distribution information about images, are extensively used in image retrieval based on colour similarity. Because image databases usually store a large number of images, sequential scanning of the database is not feasible; an index scheme is needed to speed up the similarity search. Colour histograms are high-dimensional data, usually no less than 64 dimensions. The current multi-dimensional index structures features, however, are not efficient for high-dimensional data. Indexing directly on colour histograms, therefore, is not feasible.
In this research, we concentrate on developing index schemes to support efficient image retrieval based on colour histograms. Our study shows that the proposed new index structures can handle high dimensionality of multi-dimensional data for colour histograms and help for speeding up the image similarity search. It will provide much improvement for image data retrievals.
|
|---|
