Contributions to the Bayesian Approach to Multi-View Stereo
Contributions à l'approche bayésienne pour la stéréovision multi-vues
Résumé
Multi-view stereo is the problem of recovering the shape of objects from multiple images taken from different but known camera positions. It is an inverse problem where we want to find the cause (the object) given the effect (the images). From a Bayesian perspective, the solution would be the reconstruction that best reproduces the input images while at the same time being plausible a priori. Taking this approach, in this thesis we develop generative models and methods for computing reconstructions that minimize the difference between the observed images and the images sythetized from the reconstruction.
Three models are presented. The first, represents the reconstructed scene by a set of depth maps. This gives high resolution results, but have problems at the objects boundaries. The second model represents the scene by a discreet occupancy grid, yielding to a combinatorial optimization problem, which is addressed through message passing techniques. The final model represents the scene by a smooth surface and the resulting optimization problem is solved via gradient descent surface evolution.
In either model, the main difficulty is to correctly take into account the occlusions. Modeling self-occlusions results in optimization problems that challenge current optimization techniques. In this respect, the main result of the thesis is the computation of the derivative of the reprojection error with respect to surface variations taking into account the visibility changes that occur while the surface moves. This enables the use of gradient descent techniques, and leads to surface evolutions that place the contour generators of the surface to their correct location in the images without the need of additional silhouettes or apparent contours constraints.
Three models are presented. The first, represents the reconstructed scene by a set of depth maps. This gives high resolution results, but have problems at the objects boundaries. The second model represents the scene by a discreet occupancy grid, yielding to a combinatorial optimization problem, which is addressed through message passing techniques. The final model represents the scene by a smooth surface and the resulting optimization problem is solved via gradient descent surface evolution.
In either model, the main difficulty is to correctly take into account the occlusions. Modeling self-occlusions results in optimization problems that challenge current optimization techniques. In this respect, the main result of the thesis is the computation of the derivative of the reprojection error with respect to surface variations taking into account the visibility changes that occur while the surface moves. This enables the use of gradient descent techniques, and leads to surface evolutions that place the contour generators of the surface to their correct location in the images without the need of additional silhouettes or apparent contours constraints.
La stéréovision multi-vues consiste à retrouver la forme des objets à partir de plusieurs images prises de différents points de vue connus. Ceci est un problème inverse où on cherche la cause (l'objet) alors qu'on observe l'effet (les images). Sous une optique bayésienne, la solution serait une reconstruction qui reproduise au mieux les images observées tout en restant plausible a priori. Dans cette thèse, nous présentons des modèles et des méthodes permettant de minimiser la différence entre les images observées et les images obtenues par le rendu de la reconstruction. Pour ceci, il est nécessaire de tenir compte des occultations qui on lieu lors du rendu. Le résultat principal de la thése est le calcul de la dérivée de l'erreur de reprojection par rapport aux variations de surface qui tiens en compte les changements de visibilité lors que la surface se déforme.
Loading...