Christian Bouville

Spherical Fibonacci Point Sets for Illumination Integrals

Quasi‐Monte Carlo (QMC) methods exhibit a faster convergence rate than that of classic Monte Carlo methods. This feature has made QMC prevalent in image synthesis, where it is frequently used for approximating the value of spherical integrals (e.g. illumination integral). The common approach for generating QMC sampling patterns for spherical integration is to resort to unit square low‐discrepancy sequences and map them to the hemisphere. However such an approach is suboptimal as these sequences do not account for the spherical topology and their discrepancy properties on the unit square are impaired by the spherical projection. In this paper we present a strategy for producing high‐quality QMC sampling patterns for spherical integration by resorting to spherical Fibonacci point sets. We show that these patterns, when applied to illumination integrals, are very simple to generate and consistently outperform existing approaches, both in terms of root mean square error (RMSE) and image quality. Furthermore, only a single pattern is required to produce an image, thanks to a scrambling scheme performed directly in the spherical domain.

Generating High Quality Pictures by Ray-Tracing

Ray‐casting techniques provide a very general framework in which many problems can be solved in a much easier way than with conventional methods. This is particularly true for the illumination model when a high level of realism is required. Another interesting feature of ray‐casting is its ability to display a wide class of algebraic surfaces with a minimum of approximation. Both aspects are developed in this paper where a full lighting model, based on a theoretical approach, is presented. Then, an algorithm for the display of surfaces of revolution is described.

Developments in Ray-Tracing

After a brief recall of the basic principles of ray-tracing, the traditional and distributed approaches will be presented with the related illumination models. This will include an introduction to relevant topics such as photometry and optics, Monte-Carlo integration, stochastic sampling and a discussion on implementation issues. Then, the fundamental problem of ray object intersection will be tackled. First, intersection algorithms for various type of objects will be described including simple objects (solid primitives, polygons,...) as well as more complex objects (bicubic patches, fractals, CSG modelled object,...). Then, after a discussion on bounding volumes, the case of large and complex collections of primitives is considered by examining the two basic classes of organizing structure (i.e. hierarchical data structure and space-partitioning) and their related traversal algorithms. Eventually, the parallel processing issue will be discussed.

Peer-to-peer visualization of very large 3D landscape and city models using MPEG-4

The recent availability of broadband Internet access and web-based visualization techniques is paving the way for a large scale use of 3D landscape and city models for a great variety of professional and mass market services. To make such services appealing to a large audience, these 3D models must reach a sufficient level of realism and accuracy. Many solutions are now available to automatically generate 3D models of huge urban environments. Geographic Information System (GIS) databases (i.e. terrain elevation grids, ortho photographs, buildings footprint and height) provide a good basis for generating such models at affordable cost with minimum human intervention. However, the growing size of the transmitted data will favour streaming over download-and- play. Moreover, the new market for virtual worlds is likely to be soon demanding in interoperability. The goal of this paper is to show how solutions based on hierarchical LOD models and view-dependent progressive streaming can efficiently be implemented using the MPEG-4 AFX standard in peer-to-peer networks architectures, solving both streaming and interoperability issues.

Photorealism in computer graphics

Invited Paper.- Trends in Radiosity for Image Synthesis.- 1. Efficient Ray Tracing Methods.- Incremental Ray Tracing.- Parametric Surfaces and Ray Tracing.- 2. Theory of Global Illumination Models.- A Theoretical Analysis of Global Illumination Models.- Physically Based Lighting Calculations for Computer Graphics: A Modern Perspective.- Efficient Radiosity Methods for Non-Separable Reflectance Models.- 3. Photometric Algorithms.- A Progressive Ray-Tracing-based Radiosity with General Reflectance Functions.- A Two-Pass Radiosity Method for Bezier Patches.- 4. Form Factor Calculation.- The Hemisphere Radiosity Method: A Tale of Two Algorithms.- Exploiting Coherence for Clipping and View Transformations in Radiosity Algorithms.- A Rapid Hierarchical Radiosity Algorithm for Unoccluded Environments.- Fast Radiosity by Parallelization.- 5. Physics Based Methods.- Newtons Colors: Simulating Interference Phenomena in Realistic Image Synthesis.- Light Sources in a Ray Tracing Environment.- Methods for Efficient Sampling of Arbitrarily Distributed Volume Densities.

DVFLEX: a flexible MPEG real time video codec

The MPEG-2 video coding standard offers great latitude in implementation and covers a wide variety of applications. The DVFLEX video encoder allows one to take advantage of this flexibility by making the whole coding loop including the bitstream generation, software programmable. The DVFLEX architecture is basically an array of linearly-connected processing modules. These modules are based on the MVP (TMS320C80) and include dedicated I/O ports for video, bitstream and inter-processor communications. All these I/O ports allow the processing to be performed in parallel with I/Os. The processing workload is distributed among these modules on an image region per processor basis. As for the on-chip MVP processors, tasks assignment is much more constrained because of the bitstream handling problems. Although suboptimal, the solution we describe is simple and gives satisfactory results. However, the on-chip to off-chip memory I/Os have to be carefully handled so as to keep the processors busy.

A Spherical Gaussian Framework for Bayesian Monte Carlo Rendering of Glossy Surfaces

The Monte Carlo method has proved to be very powerful to cope with global illumination problems but it remains costly in terms of sampling operations. In various applications, previous work has shown that Bayesian Monte Carlo can significantly outperform importance sampling Monte Carlo thanks to a more effective use of the prior knowledge and of the information brought by the samples set. These good results have been confirmed in the context of global illumination but strictly limited to the perfect diffuse case. Our main goal in this paper is to propose a more general Bayesian Monte Carlo solution that allows dealing with nondiffuse BRDFs thanks to a spherical Gaussian-based framework. We also propose a fast hyperparameters determination method that avoids learning the hyperparameters for each BRDF. These contributions represent two major steps toward generalizing Bayesian Monte Carlo for global illumination rendering. We show that we achieve substantial quality improvements over importance sampling at comparable computational cost.

A Bayesian Monte Carlo Approach to Global Illumination

Most Monte Carlo rendering algorithms rely on importance sampling to reduce the variance of estimates. Importance sampling is efficient when the proposal sample distribution is well‐suited to the form of the integrand but fails otherwise. The main reason is that the sample location information is not exploited. All sample values are given the same importance regardless of their proximity to one another. Two samples falling in a similar location will have equal importance whereas they are likely to contain redundant information. The Bayesian approach we propose in this paper uses both the location and value of the data to infer an integral value based on a prior probabilistic model of the integrand. The Bayesian estimate depends only on the sample values and locations, and not how these samples have been chosen. We show how this theory can be applied to the final gathering problem and present results that clearly demonstrate the benefits of Bayesian Monte Carlo.

Peer-to-Peer Visualization of Very Large 3D Landscape and City Models Using MPEG-4

The recent availability of broadband Internet access and web-based visualization techniques is paving the way for a large scale use of 3D landscape and city models for a great variety of professional and mass market services. To make such services appealing to a large audience, these 3D models must reach a sufficient level of realism and accuracy. Many solutions are now available to automatically generate 3D models of huge urban environments. Geographic Information System (GIS) databases (i.e. terrain elevation grids, ortho photographs, buildings footprint and height) provide a good basis for generating such models at affordable cost with minimum human intervention. However, the growing size of the transmitted data will favour streaming over download-and- play. Moreover, the new market for virtual worlds is likely to be soon demanding in interoperability. The goal of this paper is to show how solutions based on hierarchical LOD models and view-dependent progressive streaming can efficiently be implemented using the MPEG-4 AFX standard in peer-to-peer networks architectures, solving both streaming and interoperability issues.

A Theoretical Analysis of Global Illumination Models

This paper provides a general theoretical framework in which any previous global illumination model can be described. It gives a theoretical background which is based on radiometric quantities and on a physical approach to insure generality and consistency. A system of energy balance equations is obtained. Two approaches are possible depending on the way the radiometric properties of surfaces are characterized. Furthermore, the problem of solving this system of equations with the two-component method is addressed, and we show how we can take advantage of the characteristics of each component to select an appropriate method.