Ali, M. Ameer; Dooley, Laurence S. and Karmakar, Gour C.
PDF (Not Set)
- Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
|Google Scholar:||Look up in Google Scholar|
The segmentation results of any clustering algorithm are very sensitive to the features used in the similarity measure and the object types, which reduce the generalization capability of the algorithm. The previously developed algorithm called image segmentation using fuzzy clustering incorporating spatial information (FCSI) merged the independently segmented results generated by fuzzy clustering-based on pixel intensity and pixel location. The main disadvantages of this algorithm are that a perceptually selected threshold does not consider any semantic information and also produces unpredictable segmentation results for objects (regions) covering the entire image. This paper directly addresses these issues by introducing a new algorithm called fuzzy image segmentation using location and intensity (FSLI) by modifying the original FCSI algorithm. It considers the topological feature namely, connectivity and the similarity based on pixel intensity and surface variation. Qualitative and quantitative results confirm the considerable improvements achieved using the FSLI algorithm compared with FCSI and the fuzzy c-means (FCM) algorithm for all three alternatives, namely clustering using only pixel intensity, pixel location and a combination of the two, for a range of sample of images.
|Item Type:||Conference Item|
|Extra Information:||This paper had the Track No: 396-311|
|Academic Unit/Department:||Mathematics, Computing and Technology > Computing & Communications
Mathematics, Computing and Technology
|Interdisciplinary Research Centre:||Centre for Research in Computing (CRC)|
|Depositing User:||Laurence Dooley|
|Date Deposited:||08 Jun 2009 10:13|
|Last Modified:||24 Feb 2016 16:37|
|Share this page:|
Download history for this item
These details should be considered as only a guide to the number of downloads performed manually. Algorithmic methods have been applied in an attempt to remove automated downloads from the displayed statistics but no guarantee can be made as to the accuracy of the figures.