Fakultät Informatik

Modeling 3D Fluid Volumes Based on Appearance Transfer



Since physically-based fluid simulations are often hard to control in terms of art direction, image-based modeling of fluids is a good alternative for creating special effects. However, it is quite challenging to capture flow data of real fluid phenomena in an efficient way. Recent research explored modeling fluid volumes by appearance transfer which proved successful when applied to fire.  Now, it is interesting to check wether this technique is also successful when it comes to smoke and liquids. Additionally, the proposed method could be enhanced by implementing non-parametric texture synthesis.


The goal of this thesis is to reconstruct a fluid’s volume by using a sequence of single/pair-view images. An energy minimization problem has to be solved by applying an expectation maximum (EM)-like algorithm which makes use of a least squares method and appearance information for novel views including histogram-matching and steerable pyramids.


    • Interest / experiences in fluid simulation
    • C++, fundamental knowledge in python, advanced math skills
    • Advantageous: experiences with  mantaflow, lecture Simulation for Visual Effects
    • Please provide your CV and transcript of records





      Matthias Niessner, our new Professor from Stanford University, offers a number of interesting topics for  master theses.


      PhD positions on   Computational Fabrication and 3D Printing and  Photorealistic Rendering for Deep Learning and Online Reconstruction are available at the Computer Graphics & Visualization group.