GPU-Implementation-of-Order-Independent-Pixel-based-Texture-Synthesis

Final project for CS 759 High Performance Computing Class taken in UW-Madison

Here is documentation of Jieru Hu's ME 759 Final Project in UW-Madison 2017 Fall.

Results:

This folder contains a set of texture synthesis results from the GPU Implementation.

Sample:

This folder contains some sample texture for the program input.

Src:

This folder contains all the source code.

Timing:

This folder contains timing results stored in txt, scaling analysis plots, and a python script does all the plotting.

Instruction for Execution

This code is tested and worked on Euler in 759 class. To run the program on Euler:

cd src
sbatch ts_gpu.sh

The result texture image will be produce in src. The execution record will be produced in src/ts_gpu.out

To run the program on another machine, you would need both CUDA and OpenCV support. After obtain those libs, do the following to execute the program:

cd src
make
./ts_gpu texture1.jpg 400 3 11 2 3 //[image name] [output size] [# pyramid level] [neighborhood size] [size ratio] [#iterations per level]