Offline writer authentication
| dc.contributor.advisor | Coetzer, Johannes | en_ZA |
| dc.contributor.author | Shumba, Sandura | en_ZA |
| dc.contributor.other | Stellenbosch University. Faculty of Science. Dept. of Mathematical Sciences. Division Applied Mathematics. | en_ZA |
| dc.date.accessioned | 2021-11-22T17:57:42Z | |
| dc.date.accessioned | 2021-12-22T14:22:33Z | |
| dc.date.available | 2021-11-22T17:57:42Z | |
| dc.date.available | 2021-12-22T14:22:33Z | |
| dc.date.issued | 2021-12 | |
| dc.description | Thesis (MSc)--Stellenbosch University, 2021. | en_ZA |
| dc.description.abstract | ENGLISH ABSTRACT: In this thesis a number of systems are proposed for the purpose of offline writer authentication. A text-dependent approach is adopted, since a very specific targeted handwritten word is considered for authentication purposes. Feature extraction is facilitated by calculating a number of projections of the targeted word from different angles. Two distinct categories of systems are proposed. The first category employs template matching and is based on the computation of the Euclidean distance and a dynamic time warping (DTW) distance between corresponding feature vectors, while the second category relies on machine learning techniques, that is support vector machines (SVMs) and quadratic discriminant analysis (QDA). Within the context of the proposed machine learning-based systems, a writer-independent protocol is followed. This is achieved by employing a DTW-based dichotomy transformation which converts a feature set in feature space into a dissimilarity vector-based representation in dissimilarity space. This dichotomy transformation is followed by writer-specific dissimilarity vector normalisation which significantly improves interclass separability. The DTW-based dichotomy transformation and writer-specific dissimilarity vector normalisation are novel within the context of offline writer authentication. The systems developed in this study are evaluated on a subset of the CEDAR-LETTER data set. It is demonstrated that the proficiency of the systems developed in this study are at least on par when compared to existing systems. The most proficient SVM-based system developed in this study achieves an AUC of 93% and an equal error rate (EER) of 14.93%. | en_ZA |
| dc.description.abstract | AFRIKAANSE OPSOMMING: In hierdie tesis word ’n aantal stelsels vir die doel van vanlyn skrywerverifikasie voorgestel. ’n Teksafhanklike benadering word gevolg, aangesien ’n baie spesifieke handgeskrewe teikenwoord vir verifikasiedoeleindes beskou word. Kenmerkonttrekking word moontlik gemaak deur ’n aantal projeksies van die teikenwoord vanuit verskillende hoeke te bereken. Twee aparte kategorieë van stelsels word voorgestel. Die eerste kategorie gebruik templaatpassing en is gebaseer of die berekening van die Euklidiese afstand en ’n dinamiese tydsverbuiging (DTW) afstand tussen die ooreenstemmende kenmerkvektore, terwyl die tweede kategorie op masjienleertegnieke staatmaak, m.a.w. ondersteuningsvektormasjiene (SVMs) en kwadratiese diskriminant-analise (QDA). Binne die konteks van die voorgestelde masjienleergebaseerde stelsels word ’n skrywer-onafhanklike protokol gevolg. Dit word moontlik gemaak deur ’n DTW-gebaseerde tweeledigheidstransformasie te ontplooi wat ’n kenmerkstel in die kenmerkruimte na ’n verskilvektor-gebaseerde voorstelling in die verskilruimte omskakel. Hierdie tweeledigheidstransformasie word deur skrywerspesifieke verskilvektor-normalisasie gevolg, wat interklas-skeibaarheid aansienlik verhoog. Die DTW-gebaseerde tweeledigheidstransformasie en skrywerspesifieke verskilvektor-normalisasie is nuut binne die konteks van vanlyn skrywerverifikasie. Die stelsels wat in hierdie tesis ontwikkel is, word op ’n substel van die CEDAR-LETTER datastel geëvalueer. Dit word aangetoon dat die vaardigheid van die stelsels wat in hierdie studie ontwikkel is ten minste soortgelyk is aan dié van bestaande stelsels. Die mees vaardige SVM-gebaseerde stelsel wat in hierdie studie ontwikkel is behaal ’n AUC van 93% en ’n gelyke foutkoers (EER) van 14.93%. | af_ZA |
| dc.description.version | Masters | |
| dc.format.extent | xiii, 75 pages : illustrations | en_ZA |
| dc.identifier.uri | http://hdl.handle.net/10019.1/123807 | |
| dc.language.iso | en_ZA | en_ZA |
| dc.publisher | Stellenbosch : Stellenbosch University | en_ZA |
| dc.rights.holder | Stellenbosch University | en_ZA |
| dc.subject | Biometric identification | en_ZA |
| dc.subject | Handwriting -- Authentification | en_ZA |
| dc.subject | Template matching (Digital image processing) | en_ZA |
| dc.subject | Pattern recognition systems | en_ZA |
| dc.subject | Support vector machines | en_ZA |
| dc.subject | UCTD | |
| dc.title | Offline writer authentication | en_ZA |
| dc.type | Thesis | en_ZA |