Uncertainty Models in the Context of Biometric Authentication Systems

Eastwood, Shawn C.

Uncertainty Models in the Context of Biometric Authentication Systems

dc.contributor.advisor	Yanushkevich, Svetlana N.
dc.contributor.advisor	Shmerko, Vlad P.
dc.contributor.author	Eastwood, Shawn C.
dc.contributor.committeemember	Nielsen, John
dc.contributor.committeemember	Sesay, Abu B.
dc.contributor.committeemember	Sezer, A. Deniz
dc.contributor.committeemember	Coates, Mark J.
dc.contributor.committeemember	Dimitrov, Vassil
dc.date	2019-06
dc.date.accessioned	2019-04-26T16:08:30Z
dc.date.available	2019-04-26T16:08:30Z
dc.date.issued	2019-04-25
dc.description.abstract	This thesis focuses on developing computationally-efficient machine reasoning models. These models are based on causal graphs with various metrics of uncertainty. The application of such models is decision-making in a multi-sensor, multi-source system. In particular, we consider examples of biometric-enabled systems for human identification where false passes and false rejects are always present. Two main problems are addressed in this thesis: the potential lack of data that is needed to build an accurate model, and the computational complexity (worst case computing time) of the process of deriving conclusions from the model (uncertainty inference). To tackle the first problem, this research suggests the use of advanced models of uncertainty. These models require the development of a taxonomy of various approaches to quantifying uncertainty with the aim of being tolerant to incomplete data. Tasks related to uncertainty model design include but are not limited to: • Model training, which is the generation of uncertainty models from raw data and expert knowledge. • Major approaches to quantifying uncertainty include but are not limited to: probability distributions, fuzzy probability distributions, credal sets, probability interval distributions, Dempster-Shafer models, and Dezert-Smarandache models. To address the second problem, this work develops a platform and software to perform the calculations related to the uncertainty models in a computationally-efficient manner. Tasks related to the usage of uncertainty models include but are not limited to: • Uncertainty inference, which is the calculation of likely outcomes and uncertainty values when provided with both a model of the scenario under consideration and observed evidence. This thesis covers some approximate approaches to uncertainty inference. • Data/information fusion, which is a subset of uncertainty inference that involves the process of collecting uncertainty values or observations from various sensors, and then generating a “recommendation”. To address the problem of computational complexity, approximate approaches will be developed and utilized in this thesis. These approximate approaches are formulated with the aim of reducing the computational complexity, while maintaining a reasonable degree of accuracy. Examples of applications of the proposed theoretical developments, including risk assessment tasks in biometric-enabled systems, are provided.	en_US
dc.identifier.citation	Eastwood, S. C. (2019). Uncertainty models in the context of biometric authentication systems (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca.	en_US
dc.identifier.doi	http://dx.doi.org/10.11575/PRISM/36405
dc.identifier.uri	http://hdl.handle.net/1880/110220
dc.language.iso	eng	en_US
dc.publisher.faculty	Schulich School of Engineering	en_US
dc.publisher.institution	University of Calgary	en
dc.rights	University of Calgary graduate students retain copyright ownership and moral rights for their thesis. You may use this material in any way that is permitted by the Copyright Act or through licensing that has been assigned to the document. For uses that are not allowable under copyright legislation or licensing, you are required to seek permission.	en_US
dc.subject	Biometrics	en_US
dc.subject	Risk assessment	en_US
dc.subject	uncertainty	en_US
dc.subject	decision support	en_US
dc.subject	machine reasoning	en_US
dc.subject.classification	Engineering--Electronics and Electrical	en_US
dc.title	Uncertainty Models in the Context of Biometric Authentication Systems	en_US
dc.type	doctoral thesis	en_US
thesis.degree.discipline	Engineering – Electrical & Computer	en_US
thesis.degree.grantor	University of Calgary	en_US
thesis.degree.name	Doctor of Philosophy (PhD)	en_US
ucalgary.item.requestcopy	true