Vincent Masselus Katholieke Universiteit Leuven

Contact: Computer Graphics Research Group

Ph.D. Thesis, Katholieke Universiteit Leuven, 156 + vii p

Promotors

• Prof. Yves D. Willems, Faculty of Applied Sciences, K.U.Leuven
• Prof. Philip Dutre, Faculty of Applied Sciences, K.U.Leuven

Jury

• Prof. Marc Van Barel, Faculty of Applied Sciences, K.U.Leuven
• Prof. Luc Van Gool, Faculty of Applied Sciences, K.U.Leuven
• Prof. Philippe Bekaert, Expertise Centrum for Digitale Media, Limburgs Universitair Centrum
• Prof. Hans-Peter Seidel, Max-Planck-Institut fur Informatik, Saarbrucken, Germany
• Prof Paul E. Debevec, University of Southern California, USA

Abstract

Computer graphics allow to create increasingly more fantastic yet still believable images and the range of applications continues to grow in many areas such as the movie industry and computer games. In order to maintain the creative freedom of virtual reality and augment the realism of the resulting images, the visualized scenes have become a mixture of existing real objects and virtually created objects. This dissertation deals with the visualization of real objects with novel incident, possibly virtual illumination, which is also called relighting.

An image-based relighting technique requires only to take photographs of the object under certain conditions, and no other information, such as a geometric description or reflectance properties of the materials of the object, is required. In this work, we choose for an image-based relighting technique and out of practical considerations, the user is restricted to a fixed point of view to the object.

In the first part of this work, we design a framework for fixed viewpoint image-based relighting techniques. In this framework a function defining the transport of incident light through the object space, is captured. This function is called the reflectance field of the object. To acquire this reflectance field, a set of illumination conditions needs to be chosen and the object is photographed from the fixed viewpoint while illuminated with each of these illumination conditions. From the resulting set of images, the reflectance field of the object is constructed. Once the reflectance field is obtained, relighting the object boils down to evaluating the reflectance field with the desired incident illumination.

In the second part, we demonstrate that previously published relighting techniques fit in this framework, which allows to fairly compare these techniques. Some newly developed relighting techniques are presented in this framework as well. A first set of techniques deals with relighting real objects with restricted 2D illumination. For this kind of relighting, a very practical data acquisition method is presented, which only requires a hand-held light source, a camera and 4 white diffuse spheres. Additionally, more intelligent data processing techniques of the resulting set of images are presented and compared, allowing to construct a better reflectance field of the object. Secondly, this work also provides the first presented technique for relighting a real object with non-restricted 4D illumination, which enables to create spotlight effects or visualize the object partly in shadow. The improvements presented for relighting with restricted 2D illumination are transferred to relighting with 4D illumination as well.

---

Downloads

• PhD Thesis (Full Text, PDF, 6.5MB)
• Dutch Summary (PDF, 714KB)

Links

• Lirias

Outline

• Chapter 1: Introduction
• General introduction to the field of image-based relighting
• Overview of the contributions
• Chapter 2: General Image-based Relighting
• Light properties important for relighting
• Mathematical prerequisites
• General Image-based relighting equation
• Practical considerations: parameterization for hemispheres, acquiring incident illumination and High-Dynamic Range imaging
• Chapter 3: Relighting with 2D Illumination
• Rewriting the general relighting equation for relighting with 2D illumination
• Relighting with a Light Stage
• Relighting with a Free-form Light Stage (click to see the "Free-form Light Stage" paper page)
• Other 2D relighting techniques in the framework: Steerable functions, environment matting,...
• Chapter 4: Relighting with 4D Illumination
• Relighting with 4D illumination, rewritten for the practical framework
• Basic data acquisition speedup
• (click to see the original "Relighting with 4D Incident Light Fields" paper page
• Chapter 5: 2D Reflectance Functions
• More inteligent reconstruction of the 2D reflectance functions, using interpolation, biquadric polynomials, multilevel B-Splines and more
• Efficient representation of 2D reflectance functions, using box functions, spherical harmonics and non-linear wavelet approximations
• (click to see the "Smooth Reconstruction and Compact Representation of Reflectance Functions for Image-Based Relighting" paper page)
• Chapter 6: 4D Reflectance Functions
• Better reconstruction of 4D reflectance functions, using interpolation in between projector positions and other illumination patterns for the projector
• Efficient representation of acquired data per projector position
• Chapter 7: Conclusion

For reproduction: print the .pdf file two-sided on A4 paper. The odd side of the paper is displayed on the image below:

Presentation

The presentation (.ppt file) is available apon request. Please contact me for this.