Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/476553Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Siti Norul Huda Sheikh Abdullah, Prof. | - |
| dc.contributor.author | Baher Ayad Mousa (P53611 ) | - |
| dc.date.accessioned | 2023-10-06T09:20:55Z | - |
| dc.date.available | 2023-10-06T09:20:55Z | - |
| dc.date.issued | 2012-01-14 | - |
| dc.identifier.other | ukmvital:119379 | - |
| dc.identifier.uri | https://ptsldigital.ukm.my/jspui/handle/123456789/476553 | - |
| dc.description | Connected component labelling (CCL) is one of the fundamental operations which is usually employed in many computer vision applications. CCL algorithms assign all linked neighbouring pixels which share similar or approximate value in a binary or gray scale image correspondingly into a unique label value. However, these methods rely on complexity of the source image and thresholding process. This may affect the process of searching of the linked components as well as processing time. Therefore, this thesis proposed enhanced CCL algorithm for binary images. It comprises two steps: first is to produce the contour image by subtracting the binary image with its erosion image, and second is to compute the number of linked objects by tracing the inner and outer contour using a seed point with eight neighbouring pixels. The unique of this proposed method is during outer contour search, it skips some tracing process by comparing its binary image association. Then, the proposed method and state of the art methods are experienced and tested on a wide collection of binary and gray scale images. State of the art methods categorised into two classes: Irregular comprises one scan and contour tracing scan algorithms, and Regular comprises two and multi scan algorithms. The evaluation performances are measured based on memory and time consumption and number of labelled connected component versus thresholding methods. Experimental results show that the accuracy of all CCL algorithms depends on the used criteria for handling the gray-scale image and the used thresholding technique, while the execution time depends on the size of the testing image and the number of objects in the image. The proposed algorithm shows better performance in comparison to classic contour tracing CCL algorithm as its execution time decreased by 9.32%. The results also show that one scan algorithm is superior among all the other CCL algorithms.,Certification of Master's/Doctoral Thesis" is not available | - |
| dc.language.iso | eng | - |
| dc.publisher | UKM, Bangi | - |
| dc.relation | Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat | - |
| dc.rights | UKM | - |
| dc.subject | Connected component labelling (CCL | - |
| dc.subject | Algorithm | - |
| dc.subject | Universiti Kebangsaan Malaysia -- Dissertations | - |
| dc.title | Connected component labelling based on contour scan algorithm | - |
| dc.type | theses | - |
| dc.format.pages | 72 | - |
| dc.identifier.callno | TA1632.M696 2012 3 tesis | - |
| dc.identifier.barcode | 002575 (2012) | - |
| Appears in Collections: | Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat | |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| ukmvital_119379+SOURCE1+SOURCE1.0.PDF Restricted Access | 1.25 MB | Adobe PDF |  View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.