Paper texture identification using gabor filters and local binary pattern operators

Shihab Hamad Khaleefah (P74159)

Please use this identifier to cite or link to this item: https://ptsldigital.ukm.my/jspui/handle/123456789/476334

Full metadata record

DC Field	Value	Language
dc.contributor.advisor	Mohammad Faidzul Nasrudin, Assoc. Prof. Dr.
dc.contributor.author	Shihab Hamad Khaleefah (P74159)
dc.date.accessioned	2023-10-06T09:16:37Z	-
dc.date.available	2023-10-06T09:16:37Z	-
dc.date.issued	2016-05-25
dc.identifier.other	ukmvital:82198
dc.identifier.uri	https://ptsldigital.ukm.my/jspui/handle/123456789/476334	-
dc.description	Paper texture identification or fingerprinting is one of the well-known techniques in authenticating documents. Scanned papers are regularly involved in the fingerprinting process. A recent work suggested that the Local Binary Pattern (LBP) operators are able to handle shearing problem during image acquisition and generate decent paper fingerprints. However, the LBP operators filter out some texture information that can be useful in image description. In addition, the variance of the intensity values within a particular region of a paper's image is very low which makes a small noise or a change in the magnitude intensity of the image affects its description. Subsequently, there are number of texture analysis methods that are applied in many computer image analysis applications. The methods segment and identify images based on variations in their textures intensity or color and one of which is Gabor filter. Gabor filter has the ability to highlight edges or interested points by using different frequencies and orientations. This thesis proposes an Automated Paper Fingerprinting (APF) method that encompasses applying Gabor filters prior to the LBP operators to enhance document's fingerprint recognition accuracy. The APF is tested in a real dataset of paper texture images. The dataset contains three sub-datasets of scanned paper images with DPI capacities of 50, 100 and 150 DPIs. To check and validate the superior combinations of the Gabor filters and LBP operators, three experiments are conducted. The first experiment represents only the Gabor filters and consists of 75 tests. The second experiment represents only the LBP operators and consists of 36 tests. The third experiment represents combinations of Gabor filters and LBP operators and consists of 900 tests. Subsequently, Chi-square similarity matching algorithm is used to evaluate the extracted texture accuracy. The experiment shows that the proposed Gabor filters-LBP operators outperformed Gabor filters and LBP methods in texture identification with average accuracy of 79.58%, 97.35%, 98.65% for the 50, 100 and 150 DPI images respectively. Principally, the test results show that Gabor filters had highlighted important features from paper texture which enables the LBP to generate enhanced feature vectors. Ultimately, the research finds that applying Gabor filters prior to the LBP operators enhances papers texture identification.,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	Paper texture identification
dc.subject	Gabor filters
dc.subject	Identify images
dc.subject	Fingerprinting
dc.subject	Dissertations, Academic -- Malaysia
dc.title	Paper texture identification using gabor filters and local binary pattern operators
dc.type	theses
dc.format.pages	158
dc.identifier.barcode	002311(2016)
Appears in Collections:	Faculty of Information Science and Technology / Fakulti Teknologi dan Sains Maklumat

Files in This Item:

File	Description	Size	Format
ukmvital_82198+SOURCE1+SOURCE1.0.PDF Restricted Access		574.26 kB	Adobe PDF	View/Open

Show simple item record Recommend this item