Abstract not available

None

Thesis (M.S.)--Brigham Young University. Dept. of Computer Science, 2006. / Includes bibliographical references (p. 141-145).

Thesis (M.A.)--York University, 2006. Graduate Programme in Psychology. / Typescript. Includes bibliographical references (leaves 51-63). Also available on the Internet. MODE OF ACCESS via web browser by entering the following URL: http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&res_dat=xri:pqdiss&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&rft_dat=xri:pqdiss:MR19643

Thesis (Ph. D.)--University of California, San Diego, 1998. / Vita. Includes bibliographical references (leaves 108-117).

Thesis (M.S.)--Brigham Young University. Dept. of Computer Science, 2008. / Includes bibliographical references (p. 65-66).

No description available.

Variational methods, which have proven to be very useful to solve the ill-posed inverse problems, have been generating a lot of research interest in the image restoration problem. It transforms the restoration problem into the optimization of a well-designed variational model. While the designed model is convex, the recovered image is the global solution found by an appropriate numerical algorithm and the quality of the restored image depends on the accuracy of the designed model. Thus, a lot of efforts have been put to propose a more precise model that can produce a result with more pleasing visual quality. Besides, due to the high- dimension and the nonsmoothness of the imaging model, an efficient algorithm to find the exact solution of the variational model, is also of the research interest, since it influences the efficiency of the restoration techniques in the practical applications. In this thesis, we are interested in the designing of both the variational models for image restoration problems and the numerical algorithms to solve these models. The first objective of this thesis is to make improvements on two models for image denoising. For the multiplicative noise removal, we designed a regularizer based on the statistical property of the speckle noise, which can transform the traditional model (named by AA) into a convex one. Therefore, a global solution can be found independent of the initialization of the numerical algorithm. Moreover, the regularization term added on the AA model can help produce a sharper result. The second model is improved on the traditional ROF model by adding an edge regularization which incorporates an edge prior obtained from the observed image. Extensive experiments show that designed edge regularization has superiority to increase the texture of the recovered result and remove the staircase artifacts in the meanwhile. It is also presented that the edge regularization designed can be easily adapted into other restoration task, such as image deblurring. The second objective of this thesis is to study the numerical algorithms for a general nonsmooth imaging restoration model. As the imaging models are usually high-dimensional, the existing algorithms usually only use the first-order information of the image. Differently, a novel numerical algorithm based on the inexact Lagrangian function is proposed in this thesis, which exploits the second-order information to reach a superlinear convergence rate. Experiments show that the proposed algorithm is able to efficiently reach the solution with higher accuracy compared to the state-of-the-art algorithm

Variational methods, which have proven to be very useful to solve the ill-posed inverse problems, have been generating a lot of research interest in the image restoration problem. It transforms the restoration problem into the optimization of a well-designed variational model. While the designed model is convex, the recovered image is the global solution found by an appropriate numerical algorithm and the quality of the restored image depends on the accuracy of the designed model. Thus, a lot of efforts have been put to propose a more precise model that can produce a result with more pleasing visual quality. Besides, due to the high- dimension and the nonsmoothness of the imaging model, an efficient algorithm to find the exact solution of the variational model, is also of the research interest, since it influences the efficiency of the restoration techniques in the practical applications. In this thesis, we are interested in the designing of both the variational models for image restoration problems and the numerical algorithms to solve these models. The first objective of this thesis is to make improvements on two models for image denoising. For the multiplicative noise removal, we designed a regularizer based on the statistical property of the speckle noise, which can transform the traditional model (named by AA) into a convex one. Therefore, a global solution can be found independent of the initialization of the numerical algorithm. Moreover, the regularization term added on the AA model can help produce a sharper result. The second model is improved on the traditional ROF model by adding an edge regularization which incorporates an edge prior obtained from the observed image. Extensive experiments show that designed edge regularization has superiority to increase the texture of the recovered result and remove the staircase artifacts in the meanwhile. It is also presented that the edge regularization designed can be easily adapted into other restoration task, such as image deblurring. The second objective of this thesis is to study the numerical algorithms for a general nonsmooth imaging restoration model. As the imaging models are usually high-dimensional, the existing algorithms usually only use the first-order information of the image. Differently, a novel numerical algorithm based on the inexact Lagrangian function is proposed in this thesis, which exploits the second-order information to reach a superlinear convergence rate. Experiments show that the proposed algorithm is able to efficiently reach the solution with higher accuracy compared to the state-of-the-art algorithm

Avec l’augmentation des applications de sécurité, un sujet particulier a attiré plus d’attention que de nombreux autres domaines en reconnaissance de formes : la biométrie – la reconnaissance des personnes en fonction de leur physique et/ou de leurs traits comportementaux. Parmi ces traits, le visage se distingue par un compromis favorable entre accessibilité et fiabilité. Dans cette thèse, nous avons principalement travaillé sur une modalité relativement plus récente en reconnaissance du visage : la forme 3D. Plus avantageuse que son homologue en 2D, car intrinsèquement invariant aux changements d’illumination de pose, l’analyse 3D du visage se heurte encore à des défis majeurs tels que les déformations de surface du visage dues aux expressions ou occultations. En conséquence, une étude approfondie est effectuée sur ce sujet incluant le prétraitement des données 3D du visage, l’annotation automatique, la simulation d’altération et d’expressions en 3D, l’extraction de caractéristiques 3D et l’analyse locale de surface. / As the need and investment for security applications grow vastly, one particular topic has certainly been attracting higher attention than many other pattern recognition fields: biometrics – recognition of humans based on their physical and/or behavioral traits. Among those traits, face stands out with its favorable reconcilement between accessibility and reliability. In this doctoral dissertation, we principally investigate 3D shape which is a relatively more recent modality in face recognition. Being advantageous to its 2D counterpart for being intrinsically invariant to illumination and pose changes, 3D face recognition still encounters major challenges such as acquisition inadequacy or facial surface deformations due to expressions or occlusion. Bearing those in mind, a thorough study is carried out on this topic, including pre-processing of 3D facial data, automatic landmarking, facial expression and alteration simulations in 3D, 3D feature extraction and 3D regional surface analysis.