This paper proposes a method for searching in an image database using a query image that is similar to the intended target. The query image may be a scan or a monochrome of the image to be retrieved. Our searching algorithm makes use of Haar wavelet standard decompositions of the query and database images. To realize a fast and accurate pattern matching, the high frequencies obtained by such decompositions are changed into spectrum images using a discrete Fourier transform. These spectrum images are invariable for the shift of images. In our simulations, the resulting algorithm is applied to a search of a target image in a database consisting of 100 images. Our experimental results illustrate an extreamly high success rate for 100 scanned and 100 monochrome query images.
|ジャーナル||Research Reports on Information Science and Electrical Engineering of Kyushu University|
|出版ステータス||出版済み - 9 1997|
All Science Journal Classification (ASJC) codes
- Computer Science(all)
- Electrical and Electronic Engineering