Digital Image Segmentation Using SOM Network
Image segmentation is one of the important stages in computer vision which is necessary for various applications such as robot control and identification of military targets, as will as image analysis of remote sensing applications. In this paper the segmentation is implemented using k-means algorit...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | Arabic |
| Published: |
Mosul University
2010-06-01
|
| Series: | Al-Rafidain Journal of Computer Sciences and Mathematics |
| Subjects: | |
| Online Access: | https://csmj.mosuljournals.com/article_163868_45d5c7a3a9886ef3f930dacbea4f4e01.pdf |
| _version_ | 1811265712940908544 |
|---|---|
| author | Amera Badran |
| author_facet | Amera Badran |
| author_sort | Amera Badran |
| collection | DOAJ |
| description | Image segmentation is one of the important stages in computer vision which is necessary for various applications such as robot control and identification of military targets, as will as image analysis of remote sensing applications.
In this paper the segmentation is implemented using k-means algorithm and minimum distance with and without SOM. Segmentation with SOM is done via many stages. In the first stage initialization and reading of image is done as well as type identification and normalization.
In the second stage the neural network SOM is implemented on the resultant image to extract its main colors. In the final stage image segmentation is done by clustering method using k-means algorithm with minimum distance. Segmentation is implemented by the following steps:-
v Image is segmented into two parts using two clusters centers.
v Calculation of a suggested quality factor to test segmentation quality for that number of clusters.
v Increment number of a clusters by one, calculate a new quality factor and compare it with the previous segmentation quality factor. Iterate this until the quality factor degrades and consider the previous classification as the right one.
v When fixing right clusters centers, a new image is created by substitution of image pixel with cluster center value that is nearest to the pixel value and then displaying and saving the final image.
Finally comparison is done between the four cases of results. It has been shown from result that the use of SOM with k-means & Minimum distance algorithm in feasible, since it is depends on the variation of objects components of image. |
| first_indexed | 2024-04-12T20:27:46Z |
| format | Article |
| id | doaj.art-ebc368d1abf34a57b49a782f54a7638d |
| institution | Directory Open Access Journal |
| issn | 1815-4816 2311-7990 |
| language | Arabic |
| last_indexed | 2024-04-12T20:27:46Z |
| publishDate | 2010-06-01 |
| publisher | Mosul University |
| record_format | Article |
| series | Al-Rafidain Journal of Computer Sciences and Mathematics |
| spelling | doaj.art-ebc368d1abf34a57b49a782f54a7638d2022-12-22T03:17:49ZaraMosul UniversityAl-Rafidain Journal of Computer Sciences and Mathematics1815-48162311-79902010-06-017113515110.33899/csmj.2010.163868163868Digital Image Segmentation Using SOM NetworkAmera Badran0College of Computer Sciences and Mathematics University of Mosul, IraqImage segmentation is one of the important stages in computer vision which is necessary for various applications such as robot control and identification of military targets, as will as image analysis of remote sensing applications. In this paper the segmentation is implemented using k-means algorithm and minimum distance with and without SOM. Segmentation with SOM is done via many stages. In the first stage initialization and reading of image is done as well as type identification and normalization. In the second stage the neural network SOM is implemented on the resultant image to extract its main colors. In the final stage image segmentation is done by clustering method using k-means algorithm with minimum distance. Segmentation is implemented by the following steps:- v Image is segmented into two parts using two clusters centers. v Calculation of a suggested quality factor to test segmentation quality for that number of clusters. v Increment number of a clusters by one, calculate a new quality factor and compare it with the previous segmentation quality factor. Iterate this until the quality factor degrades and consider the previous classification as the right one. v When fixing right clusters centers, a new image is created by substitution of image pixel with cluster center value that is nearest to the pixel value and then displaying and saving the final image. Finally comparison is done between the four cases of results. It has been shown from result that the use of SOM with k-means & Minimum distance algorithm in feasible, since it is depends on the variation of objects components of image.https://csmj.mosuljournals.com/article_163868_45d5c7a3a9886ef3f930dacbea4f4e01.pdfimage segmentationsom network |
| spellingShingle | Amera Badran Digital Image Segmentation Using SOM Network Al-Rafidain Journal of Computer Sciences and Mathematics image segmentation som network |
| title | Digital Image Segmentation Using SOM Network |
| title_full | Digital Image Segmentation Using SOM Network |
| title_fullStr | Digital Image Segmentation Using SOM Network |
| title_full_unstemmed | Digital Image Segmentation Using SOM Network |
| title_short | Digital Image Segmentation Using SOM Network |
| title_sort | digital image segmentation using som network |
| topic | image segmentation som network |
| url | https://csmj.mosuljournals.com/article_163868_45d5c7a3a9886ef3f930dacbea4f4e01.pdf |
| work_keys_str_mv | AT amerabadran digitalimagesegmentationusingsomnetwork |