| 摘要 |
A mosaic image is formed from a database of source images. More particularly, the source images are analyzed, selected and organized to produce the mosaic image. A target image is divided into tile regions, each of which is compared with individual source image portions to determine the best available matching source image by computing red, green and blue channel root-mean square error. The mosaic image is formed by positioning the respective best-matching source images at the respective tile regions.
|
