Catálogo de publicaciones - libros

According to the principles of traditional 2D animation techniques, anticipation makes an animation convincing and expressive. In this paper, we present a method to generate anticipation effects for an existing animation. The proposed method is based on the visual characteristics of anticipation, that is, “Before we go one way, first we go the other way[1].” We first analyze the rotation of each joint and the movement of the center of mass during a given action, where the anticipation effects are added. Reversing the directions of rotation and translation, we can obtain an initially guessed anticipatory pose. By means of a nonlinear optimization technique, we can obtain a consequent anticipatory pose to place the center of mass at a proper location. Finally, we can generate the anticipation effects by compositing the anticipatory pose with a given action, while considering the continuity at junction and preserving the high frequency components of the given action. Experimental results show that the proposed method can produce the anticipatory pose successfully and quickly, and generate convincing and expressive anticipation effects.

Mirage is a peculiar nature phenomenon which is caused by the atmospheric refraction and total internal reflection under special weather conditions. In this paper, we propose a novel method to model and render this phenomenon. We first establish their corresponding atmospheric temperature models. Then adhering to the physical law, we calculate the light path and intensity attenuation during its propagation. To simulate the dynamic effect of mirages, we introduce a dynamic model based on atmospheric gravity waves. By incorporating GPU acceleration into the rendering process, different types of dynamic mirages under different conditions can be realistically rendered in real time, demonstrating the formation, change, and disappear of mirages.

This paper presents a fast and reliable method to trim non-solution regions in an interval ray tracing process. The “trimming algorithm” uses interval analysis to perform rejection tests in a set of pixels simultaneously, instead of individual pixels at each time. With this approach, the presented algorithm runs faster than the traditional interval ray tracing algorithm. Also, an interval algorithm to remove aliasing in the rendering of implicit surfaces is introduced. This algorithm obtains better visualizations than the traditional point sampling. This algorithm can render thin features that would be impossible to obtain with point sampling algorithms.

The objective of metafile compositing is to retrieve multi-layered Windows Metafile command records from a picture file and translate them into a set of closed contours in a single layer that delineates a set of contiguous non-overlapping regions. Such processing is useful for a variety of engineering applications including vector graphic compression and optimization which is discussed here. Primary concerns here are the multitude of degeneracies that exist when implementing a geometric algorithm of this nature. These issues are left largely unaddressed in previous literature but can be of substantial importance when attempting to develop a robust implementation.

In digital geometry processing, it is important to preserve the intrinsic properties of 3D models in geometry editing operations. One of such intrinsic properties can be described as geometric details. For point-sampled geometry, combining the Normal Geometric Details (NGDs) and the Position Geometric Details (PGDs), a useful interactive geometry local editing method is developed. The method deforms the sample points in a region of interest by manipulating handle points. In the preprocessing step, a non-local denoising algorithm is applied to smooth the input noisy point-sampled model and as a postprocessing step, a new up-sampling and relaxation procedure is proposed to refine the deformed model. The effectiveness of the proposed method is demonstrated by examples, i.e., the editing operation can deform the model while respecting the intrinsic geometric details.

Based on artistic techniques for the creation of stained glass, we introduce a method to automatically create images in stained glass stylization of images. Our algorithm first applies segmentation, and performs region simplification to merge and simplify the segments. The system then queries a database of glass swatch images and computes an optimal matching subset based on color and texture metrics. These swatches are then mapped onto the original image and 3D rendering effects including normal mapping, translucency, lead came and refraction are applied to generate the stained glass output.

A vision-based augmented reality visual guidance system is presented. It utilises naturally occurring point features and does not require a global reference frame. Keyframes extracted from a training sequ- ence are used to provide multiple local reference frames. These keyframes are selected by minimising the uncertainties in structure recovery to find an optimal tradeoff between narrow and wide baselines.

Despite tremendous work in hair simulation a unified framework for creating realistically simulated hairstyles at interactive rates is yet not available; the main reason is that complex dynamic and optical behavior of hair are computationally expensive to simulate. To have such a framework, it is essential to find optimized solutions, especially for the various physics-based tasks, which is the main bottleneck in the simulation. In this paper, we discuss various technical advancements and achievements that have been made in formulating key techniques to handle the different challenging issues involved in simulation of hair at interactive rates. Effort has been put in all the three modules of the hair simulation – hair shape modeling, hair dynamics and hair rendering.

Modelling high-quality free-form 3D objects is a very time-consuming task. Recently, hybrid subdivision surface schemes have been proposed, allowing quads to be used in surface areas with two-directional symmetry and triangles for the more free-form regions. Until now, this kind of scheme was rather theoretical, as it was not obvious how to create optimal objects with it for e.g. practical animation purposes. We describe a sketching system for quick 3D modelling using such hybrid subdivision, starting from user-drawn 2D curves. These curves are triangularized, and a skeleton is calculated, along with the 3D object. This skeleton can be used to deform the object later on. At the same time, we try to maximize surface quality, and limit the number of faces and vertices. In cases where only little detail is desired, the resulting objects can be used directly, while the object’s structure also makes them suitable for subsequent editing and adding detail, using commercially available subdivision modellers.

Rendering the structural color of natural objects or modern industrial products in the 3D environment is not possible with RGB-based graphics platforms and software and very time consuming, even with the most efficient spectra representation based methods previously proposed. Our framework allows computing full spectra light object interactions only when it is needed, i.e. for the part of the scene that requires simulating special spectra sensitive phenomena. Achieving the rendering of complex scenes with both the full spectra and RGB light and object interactions in a ray-tracer costs only some additional fractions of seconds. To prove the efficiency of our framework, we implemented a “Multilayer Film” in a simple ray-tracer. However, the framework is convenient for any complex lighting model, including diffraction or fluorescence.