期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2009
卷号：5
期号：06
出版社：Journal of Theoretical and Applied
摘要：Image retrieval is analogous to conventional text-based querying of databases except that both the query and the retrieved data are in the form of images. For example, a query image of a mountain must retrieve images of mountains from a database of images. Such matches are done using features easily extracted from images such as colour, texture, shape and geometry. The major challenge facing the researchers is that humans measure similarities between images based on the objects found therein whereas a computer based system uses “low-level” features listed above. Humans, looking at an image of an elephant, would retrieve images of elephants from a database while a computer-based system would “look for” images containing gray regions of certain size and shape. The result is a serious mismatch between human expectation and the system performance and is called the “semantic gap” in literature
