This thesis presents the initial exploratory research into an original and novel technique to enhance performance control in animatronics. An animatronic system is defined as a 3-D electro-mechanically driven facial model that can move in certain ways, when controlled by a human performer to create the "illusion of life" for a viewer. The vital elements in this form of performance are the synchronisation of lip movements to an acoustic speech signal and the animation of emotive expressions. A novel optical sensing technique is proposed based on the hypothesis that the input of distinctive articulatory or emotive movements from the performer's face would provide a more 'natural' form of control. The principle that the movement of a minimal set of points at key positions on the face can produce sufficient control information to describe the overall action is proposed to achieve this hypothesis. A comprehensive investigation into human communication, including visual speech perception and non-verbal facial expression, to define the optimum set of key points is described. Conclusions are also drawn on the primary facial actions required for successful lip synchronisation. Both the theoretical and practical aspects of the realisation of a prototype system are described. A methodology is presented for the assessment of the sensing system and the overall objectives based on the design and construction of an animatronic face, of the same dimensions as the researcher's, to produce animation of the desired actions with similar displacements. Objective analysis is achieved through the comparison of measurements by the sensor system from the performer's key point movements and those of the animatronic model. Perceptual data is generated through the visual analysis of the animated facial movement. The results and analysis of the investigations are presented in the thesis. The thesis discusses results obtained which indicate that, given certain valid assumptions, the sensor system is capable of consistent facial motion detection. It can provide sufficient control for the animatronic model to produce a limited set of facial actions in a realistic manner. Results indicate the possibilities for improved lip synchronisation and, hence, "overall character" performance.

The ability to synthesise the effects of ageing in human faces has numerous uses from aiding the search for missing people to improving recognition algorithms and aiding surgical planning. The principal contribution of this thesis is a novel method for synthesising the visual effects of facial ageing using a training set of three-dimensional scans to train a statistical ageing model. This data-base is constructed by fitting a statistical Face Model known as a Morphable Model to a set of two dimensional photographs of a set of subjects at different age points in their lives. We verify the effectiveness of this algorithm with both quantitative and psychological evaluation. Most ageing research has concentrated on building models using two-dimensional images. This has two major shortcomings, firstly some of the information related to shape change may be lost by the projection to two-dimensions; secondly the algorithms are very sensitive to even slight variations in pose and lighting. By using standard face-fitting methods to fit a statistical face model to the image we overcome these problems by reconstructing the lost shape information, and can use a model of physical rotations and light transfer to overcome the issues of pose and rotation. We show that the three-dimensional models captured by face-fitting offer an effective method of synthesising facial ageing. The second contribution is a new algorithm for ageing a face model based on Projection to Latent Structures also known as Partial Least Squares. This method attempts to separate the training set into a set of basis vectors that best explains the shape and colour changes related to ageing from those factors within the training set that are unrelated to ageing. We show that this method is more accurate than other linear techniques at producing a face model that resembles the individual at the target age and of producing a face image of the correct perceived age. The third contribution is a careful evaluation of three well known ageing methods. We use both quantitative evaluation to determine the accuracy of the ageing method, and perceptual evaluation to determine how well the model performs in terms of perceived age increase and also identity retention. We show that linear methods more accurately capture ageing and identity information if they are trained using an individualised model, and that ageing is more accurately captured if PLS is used to train the model.

Cette thèse s’inscrit dans le cadre général de la théorie des systèmes intégrables avec bords et le développement des structures algébriques associées.D’une part, nous nous attaquons au problème de la diagonalisation de l’hamiltonien du modèle XXZ avec bords non diagonaux. Nous exhibons les deux ensembles d’états propres et valeurs propres du modèle si les paramètres de bords satisfont deux conditions.D’autre part, nous introduisons un modèle de physique statistique que nous appelons le modèle face avec un bord réfléchissant. Nous calculons exactement sa fonction de partition et nous montrons que cette dernière se représente simplement sous la forme d’un unique déterminant matriciel.Nous montrons que ces deux problèmes sont reliés par la transformation vertex-face et exhibent une structure algébrique commune, l’algèbre de réflexion dynamique. Nous nous intéressons aux aspects mathématiques de cette algèbre dans le cas elliptique général,et nous introduisons deux classes de ces représentations, la représentation de co-module d’évaluation et sa duale. Nous pensons que cette algèbre est la structure clef pour l’analyse des modèles faces avec bords. En particulier, nous montrons à l’aide de twists de Drinfel’d que leur fonction de partition se représente simplement dans le cas général. Enfin, nous tentons une ’dynamisation’ du modèle à vertex ’Half-Turn-Symmetric’,et nous décrivons sa fonction de partition en termes de représentation d’évaluation de l’algèbre de Yang-Baxter dynamique, et trouvons un ensemble de conditions la déterminantunivoquement. / This thesis is embedded in the general theory of quantum integrable models withboundaries, and the development of associated algebraic structures.We first consider the question of the diagonalization of the XXZ hamiltonian with nondiagonalboundaries. We succeed to find the two sets of eigenstates and eigenvalues of themodel if the boundaries parameters satisfy two conditions.We introduce then a statistical physics model which we refer to be the face model witha reflecting end. Moreover, we compute exactly its partition function and show that it takesthe form of a simple single matrix determinant.We show that these two problems are related through the vertex-face transformationand are solved using a common algebraic structure, the dynamical reflection algebra andits dual. We focus from a mathematical perspective on this algebra in the general ellipticcase. Both the co-module evaluation representation and its dual are introduced. We believethat these structures are the key ingredients for the analysis of face models with boundaries.In particular, using the concept of Drinfel’d twists, we show that the partition function ofthese models has a simple representation in the general case.Finally, we attempt on a ’dynamization’ of the Half-Turn-Symmetric vertexmodel. Wedescribe its partition function in terms of the evaluation representation of the dynamicalYang-Baxter algebra, and find a set of conditions that uniquely determine it.