Catálogo de publicaciones - libros

We propose a method for restoring the surface of a tooth crown so that the pose and anatomical features of the tooth will work well for chewing. The system of teeth has been modeled with a 3D statistical multi-object shape model build from 3D scans of dental cast models. The restoration is carried out using the shape model statistics in a Bayesian framework to calculate the most probable tooth crown shape(s), given the fragments of one or more neighboring and opposing tooth crowns. The modeling of and reconstruction with the multi-object shape model has been realized by extending the model with a concept of elasticity that generalizes better to new teeth. The elasticity has been calculated from the surface curvature relations within and between each tooth sample, simulating a prior knowledge of the shape variation.

Crouzon syndrome is characterised by the premature fusion of cranial sutures. Recently the first genetic Crouzon mouse model was generated. In this study, Micro CT skull scannings of wild-type mice and Crouzon mice were investigated. Using nonrigid registration, a wild-type craniofacial mouse atlas was built. The atlas was registered to all mice providing parameters controlling the deformations for each subject. Our previous PCA-based statistical deformation model on these parameters revealed only one discriminating mode of variation. Aiming at distributing the discriminating variation over more modes we built a different model using Independent Component Analysis (ICA). Here, we focus on a third method, sparse PCA (SPCA), which aims at approximating the properties of a standard PCA while introducing sparse modes of variation. The results show that SPCA outperforms both ICA and PCA with respect to the Fisher discriminant, although many similarities are found with respect to ICA.

Evolutionary computation techniques are being increasingly applied to a variety of practical and scientific problems. In this paper we present a evolutionary approach for pose estimation of a known object from one image. The method is intended to be used in pose estimation from only a few model point - image point correspondences, that is, in cases in which traditional approaches often fail.

Video surveillance becomes more and more extended in industry and often involves automatic calibration system to remain efficient. In this paper, a video-surveillance system that uses stationary-dynamic cameras devices is presented. The static camera is used to monitor a global scene. When it detects a moving object, the Pan-Tilt-Zoom (PTZ) camera is controlled to be centered on this object. We describe a method of camera-to-camera calibration, integrating zoom calibration in order to command the angles and the zoom of the PTZ camera. This method enables to take into account the intrinsic camera parameters, the 3D scene geometry and the fact that the mechanism of inexpensive camera does not fit the classical geometrical model. Finally, some experiment results attest the accuracy of the proposed solution.

Motion estimation has traditionally been approached either from a pure discrete point of view, using multi-view tensors, or from a pure continuous point of view, using optical flow. This paper builds upon a novel framework of hybrid matching constraints for motion estimation, combining the advantages of both discrete and continuous methods. We will derive both bifocal and trifocal hybrid constraints and use them together with a structure estimate based on filtering techniques. A feedback from the structure estimate will be used to further refine the motion estimate. This gives a complete iterative structure and motion estimation scheme. Its performance will be demonstrated in simulated experiments.

We present an efficient method for detecting planar bilateral symmetries under perspective projection. The method uses local affine frames (LAFs) constructed on maximally stable extremal regions or any other affine covariant regions detected in the image to dramatically improve the process of detecting symmetric objects under perspective distortion. In contrast to the previous work no Hough transform, is used. Instead, each symmetric pair of LAFs votes just once for a single axis of symmetry. The time complexity of the method is log(), where is the number of LAFs, allowing a near real-time performance. The proposed method is robust to background clutter and partial occlusion and is capable of detecting an arbitrary number of symmetries in the image.

The problem of reconstructing 3D scene features from multiple views with known camera motion and given image correspondences is considered. This is a classical and one of the most basic geometric problems in computer vision and photogrammetry. Yet, previous methods fail to guarantee optimal reconstructions - they are either plagued by local minima or rely on a non-optimal cost-function.

A common framework for the triangulation problem of points, lines and conics is presented. We define what is meant by an optimal triangulation based on statistical principles and then derive an algorithm for computing the globally optimal solution. The method for achieving the global minimum is based on convex and concave relaxations for both fractionals and monomials. The performance of the method is evaluated on real image data.

Recovering a 3-D scene from multiple 2-D views is indispensable for many computer vision applications ranging from free viewpoint video to face recognition. Ideally the recovered depth map should be dense, piecewise smooth with fine level of details, and the recovery procedure shall be robust with respect to outliers and global illumination changes. We present a novel variational approach that satisfies these needs. Our model incorporates robust penalisation in the data term and anisotropic regularisation in the smoothness term. In order to render the data term robust with respect to global illumination changes, a gradient constancy assumption is applied to logarithmically transformed input data. Focussing on translational camera motion and considering small baseline distances between the different camera positions, we reconstruct a common disparity map that allows to track image points throughout the entire sequence. Experiments on synthetic image data demonstrate the favourable performance of our novel method.

Computing high quality 3D models from multi-view stereo reconstruction is an active topic as can be seen in a recent review [15]. Most approaches make the strong assumption that the surface is Lambertian. In the case of a car, this hypothesis is not satisfied. Cars contain transparent parts and metallic surfaces that are highly reflective. To face these difficulties, we propose an approach to robustly reconstruct a 3D object in translation. Our contribution is a one-dimensional tracker that uses the vanishing point computed in a first pass. We applied it to video sequences of cars recorded from a static camera. Then, we introduce a local frame for the car and use it for creating a 3D rough model. The final result is sufficient for some applications where it is needed to estimate the size of the vehicle. This model can also be used as an initialization for more precise algorithms.

This paper describes a method of stereoscopic view generation by image-based rendering in wide outdoor environments. The stereoscopic view can be generated from an omnidirectional image sequence by a light field rendering approach which generates a novel view image from a set of images. The conventional methods of novel view generation have a problem such that the generated image is distorted because the image is composed of parts of several omnidirectional images captured at different points. To overcome this problem, we have to consider the distances between the novel viewpoint and observed real objects in the rendering process. In the proposed method, in order to reduce the image distortion, stereoscopic images are generated considering depth values estimated by dynamic programming (DP) matching using the images that are observed from different points and contain the same ray information in the real world. In experiments, stereoscopic images in wide outdoor environments are generated and displayed.