In this paper, we address the problem of tracking temporal deformations between two arbitrary densely sampled point-based surfaces. We propose an intuitive and efficient resolution to the point matching problem within two frames of a sequence. The proposed method utilizes two distinct space partition trees, one for each point cloud, which both are defined on a unique discrete space. Our method takes advantage of multi-resolution concerns, voxel adjacency relations, and a specific distance function. Experimental results obtained from both simulated and real reconstructed data sets demonstrate that the proposed method can handle efficiently the tracking process even for very large point clouds. Moreover, our method is easy to implement and very fast, which provides possibilities for real-time tracking applications.

Aircraft cockpits are advanced interfaces dedicated to the interaction and exchange of observations and commands between the pilot and the flying system. The design process of cockpits is benefiting from the use of Virtual Reality technologies: early ergonomics and layout analysis through the exploration of numerous alternatives, availability all along the cockpit life cycle of a virtual product ready for experimentation, reduced usage of costly physical mock-ups. Nevertheless, the construction of a virtual cockpit with the adequate performances is very complex. Due to the fact that the CAD based digital mock-up used for setting up the virtual cockpit is very large, one challenge is to achieve interactivity while maintaining the quality of rendering. The reduction of the information contained in the CAD database shall achieve a sufficient frame rate without degradation of the geometrical visual quality of the virtual cockpit which would alleviate the relevance of ergonomics and layout studies. This paper proposes to control the simplification process by using objective criteria based on considerations about the cockpit application and the visual performances of human beings. First, it presents the results of studies on the characteristics of the Human Visual System linked to virtual reality and visualization applications. Illustrated by first results, it establishes how to control simplifications in a rational and automatic way.

Modelling tools tend to virtual sculpturing, in which a basic object is deformed by user supplied actions. The model we present aims to be generic: whatever the geometric description of the object, we can deform it to satisfy location constraints. Our model deforms the whole space, the image of a point is a blend of deformation functions with a projection matrix which allows the satisfaction of the constraints. The user can define the extent of the deformation (i.e. the part of the object to be deformed), the shape of the deformation function to create profiles and the displacement of the constraint points to be satisfied.

Our work purpose is to elaborate a geometric and motional model for deformable solids, known only by a very small number of general crosssections images. We start from a generic 3D model of the studied solid we distort according to detected boundaries in the above mentioned images. To give the solid motion model, our work uses moving point constraints and inbetweening methods on an image sequence : we select keyframes on which we interpolate characteristic points. Then we use Coons patches to compute the inbetween frames and realize an animation? On an echocardiographic application, we obtained good results in the model precision and detected edges by image processing similar to cardiologists' hand drawn edges. Besides, the model distortion, characteristic points tracking and 2D motion simulation are encouraging, although improvements are necessary.