Caustic-Design-Project

There are two major parts in this project.

1. Caustic-Design - handles Optimal-Transport and corresponding tasks
2. Target Optimization - handles 3D Optimization

Caustic-Design

Note: This step is optional.. qmake is not fully functional in the current version. Please see the Build-Section for details.

Import in QT-Creator

• Open QT-Creator
• Choose File > Open File or Project
• Navigate to Caustic_Design folder
• Choose Caustic_Design.pro and confirm Dialog with 'Open'
  

Dependencies

Following dependencies are needed for the Caustic_Designer:

• cmake
• libqt4-dev (needs to be tested)
• libcgal-dev
• libcgal-qt4-dev (needs to be tested, too)
• libblas-dev
• liblapack-dev
• libtbb-dev
• libmetis-dev
• libsuitesparse-dev (or manually via: http://faculty.cse.tamu.edu/davis/suitesparse.html)
• liblbfgs-dev
• libtinyxml-dev

Debian dependencies as one-liner:
sudo apt-get install cmake libqt4-dev libcgal-dev libcgal-qt4-dev libblas-dev liblapack-dev libtbb-dev libmetis-dev libsuitesparse-dev liblbfgs-dev libtinyxml-dev

Build

We suggest using cmake to build the project. To do so, simple:

1. Create Build Directory (e.g. mkdir build-Caustic_Design)
2. Run cmake in the build directory (e.g. cd build-Caustic_Design && cmake ../Caustic_Design/)

You may also use qmake instead of cmake if you prefer qmake. But in the current version, suitesparse does not seem to be set correctly when installing it via apt-get.

Usage

Following sections describe the usage of the different steps for the caustic-design project.

Optimal Transport

To run the optimal transport, a source-image as well as a target-image needed. The rest of the steps are done automatically.

Loading source- and target-image

• Via code: In window.cpp uncomment the lines //open(QString("/home/p/Pictures/einstein_small.png"), false); and //open(QString("/home/p/Pictures/white_small.png"), true); and replace the strings with the path to the target image and source image, respectively.
• Via UI: File > Load Image to load target image and File > Load Source Image to load source image.

NOTE: Source- and Target-Image need to be of same ratio.

Configuration
All relevant values (amount of sites, multi-scale levels) can be configured in the config.h file.

Running To finally run the Optimal Transport:

• Algorithm > Compute Optimal Transport
• When running Gradient Descent, ensure that LEVEL_MAX is set to 1

Interpolation-> Natural Neigbors CGAL

// Assigned to: Cam
algorithm->compute interpolation to load the source image, and points (.dat) and weights (.weights) data from the OTM and run the interpolation

Target Optimization

Dependencies

Following libraries are needed:

• ceres-solver: http://ceres-solver.org/building.html
• libassimp-dev

Target Surface -> C++ (3D part)

// Assigned to: Cam
Input

• Coordinates (xR) [4]
• Target Light Direction (calculated from [4])
• Surface Mesh
• Incoming Light Direction

Output

• 3D Mesh (target surface)

Computing the surface optomization
The code is located in target-surface-optimization folder
Prerequisits:

• openGL
• GLM header librairy
• glew
• glfw
• SOIL
• assimp

Debian dependencies as one-liner:
sudo apt-get install libglew-dev libsoil-dev