How smart are current image retrieval techniques?

作者：John P. Eakins · 年份：1999 · DOI：10.1049/ic:19990883 · 被引用次数：8 · 研究领域：Image Retrieval and Classification Techniques

The use of text-based indexing for information retrieval has been well established for many years, though research into improved techniques remains active. Indexing of image data by features derived from the images themselves-generally known as content-based image retrieval (CBIR)-has a much shorter history. Originally seen as a specialist branch of image processing, CBIR has benefited greatly over the last three or four years from insights derived from fields as diverse as information retrieval, database management, and Gestalt psychology. The article considers to what extent has the field of artificial intelligence the potential to contribute to the development of CBIR. (4 pages)