Reproduction de : Thèse d'exercice : Médecine : Nancy 1 : 2007. / Titre provenant de l'écran-titre.

Thesis (Ph. D.)--University of Florida, 2001. / Title from first page of PDF file. Document formatted into pages; contains xi, 124 p.; also contains graphics. Vita. Includes bibliographical references (p. 120-123).

Thesis (M.F.A.)--Rochester Institute of Technology, 2003. / Typescript. Includes bibliographical references (leaves 21-23).

Thesis (Ph. D.)--University of Virginia, 2003. / Typescript. Includes bibliographical references (leaves 456-472). Also available online through Digital Dissertations.

Automatically determining the temporal characteristics of facial expressions has extensive application domains such as human-machine interfaces for emotion recognition, face identification, as well as medical analysis. However, many papers in the literature have not addressed the step of determining when such expressions occur. This dissertation is focused on the problem of automatically segmenting macro- and micro-expressions frames (or retrieving the expression intervals) in video sequences, without the need for training a model on a specific subset of such expressions. The proposed method exploits the non-rigid facial motion that occurs during facial expressions by modeling the strain observed during the elastic deformation of facial skin tissue. The method is capable of spotting both macro expressions which are typically associated with emotions such as happiness, sadness, anger, disgust, and surprise, and rapid micro- expressions which are typically, but not always, associated with semi-suppressed macro-expressions. Additionally, we have used this method to automatically retrieve strain maps generated from peak expressions for human identification. This dissertation also contributes a novel 3-D surface strain estimation algorithm using commodity 3-D sensors aligned with an HD camera. We demonstrate the feasibility of the method, as well as the improvements gained when using 3-D, by providing empirical and quantitative comparisons between 2-D and 3-D strain estimations.

published_or_final_version / Dentistry / Master / Master of Dental Surgery

Facial expression recognition has become an active research topic in recent years due to its applications in human computer interfaces and data-driven animation. In this thesis, we focus on the problem of how to e?ectively use domain, temporal and categorical information of facial expressions to help computer understand human emotions. Over the past decades, many techniques (such as neural networks, Gaussian processes, support vector machines, etc.) have been applied to facial expression analysis. Recently graphical models have emerged as a general framework for applying probabilistic models. They provide a natural framework for describing the generative process of facial expressions. However, these models often su?er from too many latent variables or too complex model structures, which makes learning and inference di±cult. In this thesis, we will try to analyze the deformation of facial expression by introducing some recently developed graphical models (e.g. latent topic model) or improving the recognition ability of some already widely used models (e.g. HMM). In this thesis, we develop three di?erent graphical models with di?erent representational assumptions: categories being represented by prototypes, sets of exemplars and topics in between. Our ¯rst model incorporates exemplar-based representation into graphical models. To further improve computational e±- ciency of the proposed model, we build it in a local linear subspace constructed by principal component analysis. The second model is an extension of the recently developed topic model by introducing temporal and categorical information into Latent Dirichlet Allocation model. In our discriminative temporal topic model (DTTM), temporal information is integrated by placing an asymmetric Dirichlet prior over document-topic distributions. The discriminative ability is improved by a supervised term weighting scheme. We describe the resulting DTTM in detail and show how it can be applied to facial expression recognition. Our third model is a nonparametric discriminative variation of HMM. HMM can be viewed as a prototype model, and transition parameters act as the prototype for one category. To increase the discrimination ability of HMM at both class level and state level, we introduce linear interpolation with maximum entropy (LIME) and member- ship coe±cients to HMM. Furthermore, we present a general formula for output probability estimation, which provides a way to develop new HMM. Experimental results show that the performance of some existing HMMs can be improved by integrating the proposed nonparametric kernel method and parameters adaption formula. In conclusion, this thesis develops three di?erent graphical models by (i) combining exemplar-based model with graphical models, (ii) introducing temporal and categorical information into Latent Dirichlet Allocation (LDA) topic model, and (iii) increasing the discrimination ability of HMM at both hidden state level and class level. / published_or_final_version / Computer Science / Doctoral / Doctor of Philosophy

Objectives: To report the incidence of objective and subjective neurosensory disturbance (NSD) after orthognathic surgery in a major orthognathic centre in Hong Kong, and to investigate the risk factors that contributed to the incidence of NSD after orthognathic surgery. Materials and Methods: A retrospective cross-sectional study on NSD after orthognathic surgery in a local major orthognathic centre. Patients who had bimaxillary orthognathic surgery reviewed at post-operative 6 months, 12 months or 24 months were recruited to undergo a neurosensory test with subjective and 3 objective assessments. Possible risk factors of NSD including subjects’ age and gender, surgical procedures and surgeons’ experience were analyzed. Results: 238 patients with 476 sides each of maxillary and mandibular procedures were recruited. The incidences of subjective NSD after maxillary procedures were 16.2%, 13% and 9.8% at post-operative 6 months, 12 months and 24 months, respectively; the incidences of subjective NSD after mandibular procedures were 35.4%, 36.6% and 34.6% at post-operative 6 months, 12 months and 24 months, respectively. Objective neurosensory tests showed general reduced sensitivity in subjects with subjective NSD. Increased age was found to be a significant risk factor of NSD after orthognathic surgery at short term (at 6 months and 12 months) but not at 24 months. SSO has a significantly higher risk of NSD when compared to VSSO. SSO in combination with anterior mandibular surgery has a higher risk of NSD when compared to VSSO in combination with anterior mandibular surgery or anterior mandibular surgery alone. Gender of patients a nd surgeons’ experience were not found to be risk factors of NSD after orthognathic surgery. Conclusion: The incidence of NSD after maxillary and mandibular orthognathic procedures at post-operative 6 months, 12 months and 24 months was reported. Increased age was identified as a risk factor of short term post-operative NSD but not in long term (24 months or more). Specific mandibular procedures were related to higher incidence of NSD after orthognathic surgery. / published_or_final_version / Dental Surgery / Master / Master of Dental Surgery

The relationship between infants’ neural processing of and visual preferences for attractive and unattractive faces was investigated through the integration of event-related potential and preferential looking methods. Six-month-olds viewed color images of female faces previously rated by adults for attractiveness. The faces were presented in contrasting pairs of attractiveness (attractive/unattractive) for 1.5-second durations. The results showed that compared to attractive faces, unattractive faces elicited larger N290 amplitudes at left hemisphere electrode sites (PO9) and smaller P400 amplitudes at electrode sites across both hemispheres (PO9 and PO10). There were no significant differences between infants’ overall looking times based on attractiveness, however, a significant relationship was found between amplitude and trial looking time; larger N290 amplitudes were associated with longer trial looking times. The results suggest that compared to attractive faces, unattractive faces require greater cognitive resources and longer initial attention for visual processing. / text

Deformable modeling of facial soft tissues have found use in application domains such as human-machine interaction for facial expression recognition. More recently, such modeling techniques have been used for tasks like age estimation and person identification. This dissertation is focused on development of novel image analysis algorithms to follow facial strain patterns observed through video recording of faces in expressions. Specifically, we use the strain pattern extracted from non-rigid facial motion as a simplified and adequate way to characterize the underlying material properties of facial soft tissues. Such an approach has several unique features. Strain pattern instead of the image intensity is used as a classification feature. Strain is related to biomechanical properties of facial tissues that are distinct for each individual. Strain pattern is less sensitive to illumination differences (between enrolled and query sequences) and face camouflage because the strain pattern of a face remains stable as long as reliable facial deformations are captured. A finite element modeling based method enforces regularization which mitigates issues (such as temporal matching and noise sensitivity) related to automatic motion estimation. Therefore, the computational strategy is accurate and robust. Images or videos of facial deformations are acquired with video camera and without special imaging equipment. Experiments using range images on a dataset consisting of 50 subjects provide the necessary proof of concept that strain maps indeed have a discriminative value. On a video dataset containing 60 subjects undergoing a particular facial expression, experimental results using the computational strategy presented in this work emphasize the discriminatory and stability properties of strain maps across adverse data conditions (shadow lighting and face camouflage). Such properties make it a promising feature for image analysis tasks that can benefit from such auxiliary information about the human face. Strain maps add a new dimension in our abilities to characterize a human face. It also fosters newer ways to capture facial dynamics from video which, if exploited efficiently, can lead to an improved performance in tasks involving the human face. In a subsequent effort, we model the material constants (Young's modulus) of the skin in sub-regions of the face from the motion observed in multiple facial expressions. On a public database consisting of 40 subjects undergoing some set of facial motions, we present an expression invariant strategy to matching faces using the Young's modulus of the skin. Such an efficient way of describing underlying material properties from the displacements observed in video has an important application in deformable modeling of physical objects which are usually gauged by their simplicity and adequacy. The contributions through this work will have an impact on the broader vision community because of its highly novel approaches to the long-standing problem of motion analysis of elastic objects. In addition, the value is the cross disciplinary nature and its focus on applying image analysis algorithms to the rather difficult and important problem of material property characterization of facial soft tissues and their applications. We believe this research provides a special opportunity for the utilization of video processing to enhance our abilities to make unique discoveries through the facial dynamics inherent in video.