|
|
Optimization Techniques in Computer Graphics and Appearance Fabrication
Rittig, Tobias ; Wilkie, Alexander (advisor) ; Stamminger, Marc (referee) ; Lensch, Hendrik (referee)
Tobias Rittig Optimization Techniques in Computer Graphics and Appearance Fabrication Optimization, the process of improving an intermediate solution, has been applied in numerous fields of computer science and beyond. Visual computing has throughout been at the forefront of developing new techniques and applying them to synthesize or analyze visual reality. At the latest, the recent boom of deep learning has given attention to large- scale differentiable computation and the application of gradient-descent optimization. This thesis spans parts of this development in four shared first-author publications (three journal, one conference) and three co-authored journal papers. It shows how optimization algorithms are used in two distinct fields of computer graphics. First, the focus is on the emerging field of Appearance Fabrication using full-color 3D printing. We show the benefits of an iterative optimization loop on the sharpness and color accuracy of translucent printouts. Such a loop, consisting of a forward pre- diction and backward refinement, can be composed of various building blocks developed throughout the thesis. The forward prediction can be driven by accurate Monte Carlo path tracing or an approximate neural rendering solution. The backward refinement can rely on a heuristic or a gradient-descent...
|
|
|
Affordable optical measurement methods for predictive rendering
Iser, Tomáš ; Wilkie, Alexander (advisor) ; Gkioulekas, Ioannis (referee) ; Babaei, Vahid (referee)
Predictive rendering, a part of computer graphics, is based on the light transport equation and focuses on accurately predicting the appearance of objects and materials under various conditions. A variety of problems can be formulated as appearance prediction: from generating photorealistic images to enhancing color 3D printing. The accuracy relies on the materials' optical properties, which must either be estimated from first principles, or measured with expensive and sophisticated optical devices. Could we obtain these properties in an efficient and affordable way optimized for predictive rendering? To answer the question, this thesis bridges the boundary between computer graphics and optics. We develop simpler and more affordable methods for measuring optical properties with a focus on color accuracy, thus making predictive rendering more accessible. We aim at two types of materials that are both ubiquitous but usually neglected because of their complex characteristics: translucent materials and fluorescent materials. For each, we present a separate measurement approach that only uses low-cost optical components, yet has a high spectral resolution for color-accurate applications. Our first method is motivated by measuring translucent inks, which is required for accurate full-color 3D-printing algorithms....
|
|
|
Advanced Monte Carlo methods in Image Synthesis
Vévoda, Petr ; Wilkie, Alexander (advisor)
Monte Carlo (MC) integration is an essential tool in many fields of science. In image synthesis, it has enabled photorealistic rendering results via physically based light trans- port simulation. However, an inherent problem of MC integration is variance causing noise in rendered images. This thesis presents three methods, each taking a different approach to variance reduction in rendering. The first approach focuses on improving the sampling technique. An adaptive solution is proposed for unbiased direct illumination sampling, employing Bayesian regression and a novel statistical model of direct illumination to achieve robustness. This method was integrated into a production renderer, demonstrating both its theoretical soundness and practical utility. The second approach explores the combination of multiple sampling techniques via multiple importance sampling (MIS). Optimal weighting functions are derived, proving to minimize the variance of MIS estimators. The new weights outperform all common heuristics and provide novel design considerations for selecting appropriate sampling tech- niques in integration problems. Finally, the third approach involves pre-computation to handle challenging scenarios effectively. Pre-computed reference images of a clear sky are used to create a high- quality fitted model,...
|
|
|
Advanced Monte Carlo methods in Image Synthesis
Vévoda, Petr ; Wilkie, Alexander (advisor) ; Eisemann, Elmar (referee) ; Schudeiske, Johannes (referee)
Monte Carlo (MC) integration is an essential tool in many fields of science. In image synthesis, it has enabled photorealistic rendering results via physically based light trans- port simulation. However, an inherent problem of MC integration is variance causing noise in rendered images. This thesis presents three methods, each taking a different approach to variance reduction in rendering. The first approach focuses on improving the sampling technique. An adaptive solution is proposed for unbiased direct illumination sampling, employing Bayesian regression and a novel statistical model of direct illumination to achieve robustness. This method was integrated into a production renderer, demonstrating both its theoretical soundness and practical utility. The second approach explores the combination of multiple sampling techniques via multiple importance sampling (MIS). Optimal weighting functions are derived, proving to minimize the variance of MIS estimators. The new weights outperform all common heuristics and provide novel design considerations for selecting appropriate sampling tech- niques in integration problems. Finally, the third approach involves pre-computation to handle challenging scenarios effectively. Pre-computed reference images of a clear sky are used to create a high- quality fitted model,...
|
|
|
Appearance matching and fabrication using differentiable material models
Nindel, Thomas Klaus ; Wilkie, Alexander (advisor) ; Chalmers, Alan (referee) ; Magnor, Marcus (referee)
Computing derivatives of code - with code - is one of the key enablers of the machine learning revolution. In computer graphics, automatic differentiation allows to solve in- verse rendering problems. There, parameters such as an objects reflectance, position, or the scattering- and absorption coefficients of a volume, are recovered from one or several input images. In this work, we consider appearance matching and fabrication problems, that can be cast as instances of inverse rendering problems. While gradient-based opti- mization that is enabled by differentiable programs has the potential to yield very good results, it requires proper handling - differentiable rendering is not a shotgun-type prob- lem solver. We discuss both theoretical concepts and the practical implementation of differentiable rendering algorithms, and show how they connect to different appearance matching problems. 1
|
|
|
Efficient and Expressive Microfacet Models
Atanasov, Asen ; Wilkie, Alexander (advisor) ; Yan, Lingqi (referee) ; Ribardière, Mickaël (referee)
Title: Efficient and Expressive Microfacet Models Author: Asen Atanasov Department: Department of Software and Computer Science Education Supervisor: doc. Dr. Alexander Wilkie, Department of Software and Computer Science Education Abstract: In realistic appearance modeling, rough surfaces that have micro- scopic details are described using so-called microfacet models. These include analytical models that statistically define a physically-based microsurface. Such models are extensively used in practice because they are inexpensive to compute and offer considerable flexibility in terms of appearance control. Also, small but visible surface features can easily be added to them through the use of a normal map. However, there are still areas in which this general type of model can be improved: important features like anisotropy control sometimes lack analytic solutions, and the efficient rendering of normal maps requires accurate and general filtering algorithms. We advance the state of the art with regard to such models in these areas: we derive analytic anisotropic models, reformulate the filtering problem and propose an efficient filtering algorithm based on a novel filtering data structure. Specifically, we derive a general result in microfacet theory: given an arbitrary microsurface defined via standard...
|
|
|
An Efficient Load-balancing Image Sampler for Path Tracing
Kubíček, Tomáš ; Wilkie, Alexander (advisor) ; Nindel, Thomas Klaus (referee)
In this thesis, we go through the process of improving a research renderer called ART. We focus primarily on the image sampler. We analyse the shortcomings of the stochastic sampler that was present before this thesis was written. Those being the huge memory consumption, lack of real-time visualization, and a limited communication channel between the renderer and the application that is supposed to control it. We then design two samplers that could serve as a replacement. We choose the one which seems to be better suited for the task. We then run a series of tests which serve to finetune parameters of the new sampler and also serve as a comparison between the new and the old sampler. By using the data from the tests we show that we have reduced the memory foot-print while maintaining the speed at mostly the same level as before while enabling working with greater number of cores. 1
|
|
|
Best View of 3D Scene
Jankovič, Ivan ; Pelikán, Josef (advisor) ; Wilkie, Alexander (referee)
The computation of good viewpoints is useful in many fields: exploration of virtual worlds, global illumination, robot motion, radiosity, imagebased modeling, etc. The main goal of the thesis was to implement several methods for computing the optimal viewpoint of a 3D scene. Two methods have been implemented - Heuristic Measure and Viewpoint Entropy. In addition, a modification of the first method, Weighted Heuristic Measure, was created and tested to find the values that give the best results. For the purpose of method evaluation, a simple application was developed, which loads a 3D scene and runs the implemented methods.
|
|
|
Procedurální generování měst
Beneš, Jan ; Wilkie, Alexander (advisor) ; Křivánek, Jaroslav (referee)
This thesis proposes a novel method for procedural generation of urban street networks. A simple traffic simulation, based on observations made about the social-geographic nature of cities, that is able to drive the temporal evolution of city road networks by creating new major roads, as well as a modified generation algorithm for minor roads, are described. The presented method is highly automated.
|
|
|
Wang Tiles for Frequency-Based Synthesis of Ocean Surfaces
Vámošová, Mária ; Wilkie, Alexander (advisor) ; Krajíček, Václav (referee)
Frequency-based ocean synthesis is a class of methods for generating ocean waves in deep water using physically based ocean spectra provided by oceanographic research. The output of these methods is a heightfield which is periodic and can be used to tile a larger area. However, when using a larger number of tiles, a repetition pattern becomes apparent. Wang tiling is a method of tiling the plane non- periodically by using only a small set of input tiles with given constraints. The aim of this thesis is to create a set of Wang Tiles, each of which represents an ocean heightfield synthesised by a frequency- based method, so one can create large ocean scenes without apparent periodicity in the tiling pattern. Powered by TCPDF (www.tcpdf.org)
|
guest ::
