Catálogo de publicaciones - libros

In hand drawn animation, some of animations are created at very low frame rates because of limited budget. A simple technique for increasing frame rates is animation tweening. In fact, most animation tweening applications process only vector objects. And, the important requirement for animation tweening application is user-defined motion trajectories. In this paper we propose a method to fully automate the tweening process for bit-mapped key frames. Our method is based on motion-compensated frame interpolation (MCFI) which is widely used in compression in digital video coding. Motion estimation technique used in this framework is pyramidal iterative Lucas-Kanade optical flow because the common optical flow algorithms (Barron et al., ; Barron et al., 1994; Camus, 1995; Liu et al., 1998; Benoit and Ferrie, 1996) can not handle large pixel motions. We propose a newly developed motion-compensated frame interpolation in cooperative with one-to-many relation optical flow field and valid interpolated backward difference vectors. One-to-many relation optical flow field can solve the uncovered regions in case that the objects have large different in shape and size. Valid interpolated backward difference vectors can solve the uncovered regions on the border areas of the image when some objects have first occurrence on the latter key frame. The proposed method can perform better results than classical optical flow based motion-compensated frame interpolation especially on uncovered regions.

Simulation of dynamic motion of soft tissues is one of the most important topic in nowadays computer animation of human body. So far various elastically deformable models were used for this purpose. We present the extension of model proposed in with volume preserving constraint. We provide also evaluation of the constraint accuracy on two simulation experiments. Furthermore we add model of active and passive force generated by musculo-tendon unit using Zajac’s bio-mechanical model. We use Finite Difference Method for resolution of continuous partial differential equations of internal elastic forces of the model in space. We integrate resulting ordinary differential equations of motion using various explicit and implicit methods.

We present the main results of a complete study concerning a small metallic statuette. A collaboration involving metallurgy, optics, computer graphics, archaeology and engineering specialists was needed. The virtual optical aspect computations obtained by spectral simulation is founded on optical measurements of metallic samples wether smooth or rough. The scientific study concerns a small statuette representing a horse originating from China and dated of the “Période des Royaumes combattants” (4th-2nd century BC). The obtained results are relative to reasonable hypotheses on metallurgical knowledge and practice of moulding with bronze. Some pictures both captured from the real object or spectrally computed with the digitized samples are presented.

We present methods to visually stratify overlapping markup for improved visualization of complex XML-based annotations of digital images. The problem arises in our development of editorial tools for creating image-based electronic editions, where overlapping and dense annotations make traditional markup representations cluttered, infeasible, or both. We discuss various techniques and formulate a set of constraints for efficient visualization of such markup. On the abstract level, the problem translates into visualizing data for nested and intersecting sub-polylines in a straight-segment polygonal curve. Using methods from computational geometry, such as power diagrams, we provide a complete and efficient solution to the problem.

Tensors represent a natural means for describing many physical phenomena. Therefore, there is a need to develop methods that will assist in analyzing such data. Despite some calls for generality, we concentrate on quite a specific application because general methods designed for visualizing information contained in second order tensor fields may fail to provide users with some of the features that could be of high importance for them. In the following paragraphs, our approach focused on second order tensor field datasets describing the state of stress in material will be discussed. It is based on the visualization of directional stresses using glyphs as well as surface extraction.

In this paper we describe a vehicle borne data acquisition system for urban environments and associated 3D data management and interactive rendering software. The data acquisition system is capable of acquire 3D data from urban areas with centimetre resolution including automatic capturing of colour. The system includes a management and interactive rendering software which is designed to cope with the huge quantities of data generated by the acquisition system. It uses out-of-core pre-processing to transform data into octrees. Real-time interactive rendering is achieved by using novel techniques such as front-to-back octree traversal, occlusion query and speculative prefetching. The paper presents the results of the described techniques applied to large public areas including the City Centre of Verona, Italy.

This paper presents an approach based on evolutionary computations to the IFS inverse problem. A method using variable number of mappings is proposed. Some experimental results are also shown.

This paper presents a possibility of improving the Eigenfaces method for face recognition by applying masks and eigenvectors weights. An idea of error function is introduced, which minimization optimizes the mask and weights and improves the recognition results.

This paper presents an algorithm for the segmentation of small images which combines edge-based and colour-based approaches to image segmentation. The idea is to trace continuous lines of equal intensity through an image. The starting points for these level curves are determined in an edge detection step. In a postprocessing step the gradient information along a level curve is analysed to assert that it always follows object boundaries. The segmentation algorithm works with subpixel accuracy and unites the good adaption to local image structures from the edge-based and the continuousness of segment borders from the color-based approach.

Texture segmentation methods are widely used in many image analysis tasks. However, accuracy of these methods is seldom discussed in the literature. This paper evaluates two texture segmentation techniques: based on multilayer perceptron and network of synchronized oscillators. Their accuracy of estimation of artificial image object parameters is assessed. The evaluation is performed for optical images, acquired using scanner, digital and TV camera. Evaluation results are presented along with discussion and conclusions.