Learn More
We address the problem of deformable shape and motion recovery from point correspondences in multiple perspective images. We use the low-rank shape model, i.e. the 3D shape is represented as a linear combination of unknown shape bases. We propose a new way of looking at the low-rank shape model. Instead of considering it as a whole, we assume a(More)
This paper deals with shading and AAMs. Shading is created by lighting change. It can be of two types: self- shading and external shading. The effect of self-shading can be explicitly learned and handled by AAMs. This is not however possible for external shading, which is usually dealt with by robustifying the cost function. We take a different approach: we(More)
An Active Appearance Model (AAM) is a variable shape and appearance model built from annotated training images. It has been largely used to synthesize or fit face images. Person-independent face AAM fitting is a challenging open issue. For standard AAMs, fitting a face image for an individual which is not in the training set is often limited in accuracy,(More)
Automatic extraction of facial feature deformations (either due to identity change or expression) is a challenging task and could be the base of a facial expression interpretation system. We use Active Appearance Models and the simultaneous inverse compositional algorithm to extract facial deformations as a starting point and propose a modified version(More)
Fitting a single generic AAM on an unseen face (that is not in the training set) under any pose and expression is very difficult. The variability of the data is so high that the fitting process usually gets stuck into one of the numerous local minima. We show that a solution to this problem consists to separate the variability sources. We build a pool of(More)
Le recalage direct et agrégatif d'images d'une séquence vi-déo consiste à agréger de manière incrémentale les images en déterminant pour chaque nouvelle image une transformation telle que la différence d'intensité des pixels soit mi-nimisée. Un des défauts majeurs de cette approche directe est la nécessité de disposer d'un a priori sur la zone de(More)
An Active Appearance Model (AAM) is a variable shape and appearance model built from annotated training images. It has been largely used to synthesize or fit face images. Person-independent face AAM fitting is a challenging open issue. For standard AAMs, fitting a face image for an individual which is not in the training set is often limited in accuracy,(More)
Le travail présenté ici s'inscrit dans le domaine de l'analyse assistée d'expressions faciales observées sur des séquences vidéos de langue des signes. Dans ce contexte, le visage du signeur est fréquemment occulté, principalement par ses mains. Nous utilisons un mo-dèle à apparence active (AAM) pour modéliser le vi-sage et ses déformations expressives. Un(More)
In this paper a methodology is proposed to segment beard and glasses on frontal face images. Active Appearance Model (AAM) generative power is used to find the equivalent face without beard nor glasses. These artefacts can be detected and then segmented by difference between the original face image and its free-from-artefact equivalent. We replace the(More)