|
|
|
| Basic information |
|
Two-Step Region Segmentation Method for Color Images Using Statistics of Location and Color Information : Region Segmentation by Uniform Color and Concentrated but not all Connected Pixels
Fumihiko MORI, Hiromitsu YAMADA, Makoto MIZUNO
The Institute of Electronics, Information and Communication Engineers （IEICE）
Presence/absence of peer review |
 |
Batu
Presence/absence of request |
|
Batu(Ordinary Paper)
Language of publication used |
|
Japanese
We have ever proposed a method of extracting regions from a binary image quickly based on not all connectivity-rule using statistics of location. In this paper, we propose a two-step region segmentation method for color images using statistics of brightness and color information in addition to location. The first-step of region segmentation is done for an image composed of blocks like 5x5, where the standard deviation of color information of block must be smaller than a threshold. The second-step of region segmentation is done for each pixel based on the result of the first-step.
|
