Offline Signature Verification

dc.contributor.guideSawant R R
dc.coverage.spatialSignature Verification
dc.creator.researcherNarwade Pradeep Narayan
dc.date.accessioned2020-02-26T11:50:58Z
dc.date.available2020-02-26T11:50:58Z
dc.date.awarded08/02/2020
dc.date.completed10/06/2019
dc.date.registered10/10/2013
dc.description.abstractBiometrics has always been an integral part of human identification and verification, newlineand Offline handwritten signature is a convincing evidence form of biometrics for newlineverification. It is challenging task due to the allowed variations in the individual s newlinesignatures (the skilled forgeries may have more similarity than genuine signatures) newlineand the time variant behavior of individual which affects the signatures shape. Also, newlinethe dynamic information of the signature is lost in offline signature acquisition process newlinewhich makes the offline signature verification more difficult. The main goal of newlinethe thesis is to design the algorithm to differentiate forged signatures and genuine newlinesignatures. newlineTo address the above difficulty, firstly, we proposed a novel approach to identify newlinethe correspondence between pixels of different signatures using an adaptive newlineweighted combination of shape context distance and Euclidean distance. These correspondences newlineare then further used for the transformation of query signature plane newlineto reference signature plane using thin plate spline transformation. The distances between newlinesignatures are computed using plane transformation, a shape descriptor, and newlinethe farness between matched pixels. The computed distances are fed to the Support newlineVector Machine (SVM) classifier to determine the merit of genuineness. We achieved newline89.58% accuracy using this proposed method on GPDS synthetic signature database. newlineSecondly, the shape s structure around the pixels of signature are analyzed and accurately newlinedescribed for offline signature verification. The curve angle is an important newlinedescription of shape at the corresponding pixel. In traditional curve angle (tangent newlineangle), the pixels which are far away from the center pixel have more impact on newlinecurve angle, but the pixels which are nearer to the center pixels are more important. newlineTo address this difficulty, a new curve angle i.e. Gaussian Weighting Based Tangent newlineAngle (GWBTA) is proposed. This proposed GWBTA is used to construct a new shape newlinedescriptor, i.e. cylindrical shape co
dc.description.noteBibliography
dc.format.accompanyingmaterialNone
dc.format.dimensions
dc.format.extent116p
dc.identifier.urihttp://hdl.handle.net/10603/278806
dc.languageEnglish
dc.publisher.institutionDepartment of Electronics and Telecommunication Engineering
dc.publisher.placeNanded
dc.publisher.universitySwami Ramanand Teerth Marathwada University
dc.relation115b
dc.rightsuniversity
dc.source.universityUniversity
dc.subject.keywordElectronics and telecommunication
dc.titleOffline Signature Verification
dc.title.alternative
dc.type.degreePh.D.

Files

Original bundle

Now showing 1 - 5 of 16
Loading...
Thumbnail Image
Name:
01_title.pdf
Size:
53.82 KB
Format:
Adobe Portable Document Format
Description:
Attached File
Loading...
Thumbnail Image
Name:
02_certificate.pdf
Size:
39.63 KB
Format:
Adobe Portable Document Format
Loading...
Thumbnail Image
Name:
03_abstract.pdf
Size:
42.12 KB
Format:
Adobe Portable Document Format
Loading...
Thumbnail Image
Name:
04_declaration.pdf
Size:
39.44 KB
Format:
Adobe Portable Document Format
Loading...
Thumbnail Image
Name:
05_acknowledgement.pdf
Size:
40.27 KB
Format:
Adobe Portable Document Format

License bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
license.txt
Size:
1.79 KB
Format:
Plain Text
Description: