The project's goal is the develop a GUI application where the user can transfer the artistic style from well known pictures painted by hungarian artists to their chosen images\videos.
- Python 3.5 or above: https://www.python.org/downloads/ (for Windows recommended to use Anaconda: https://www.continuum.io/downloads)
- Tensorflow: https://www.tensorflow.org/
- opencv 3.0.0 or above: https://github.com/opencv/opencv
- PyQt5: https://www.riverbankcomputing.com/software/pyqt/download5, https://www.qt.io/qt5-9/
- SciPy: https://www.scipy.org/
It is strongly recommended to use GPU for machine learning. In order of this, the following dependencies have to be met:
- CUDA compatible video card with CUDA toolkid installed: https://developer.nvidia.com/cuda-downloads
- cudnn: https://developer.nvidia.com/cudnn
- matplotlib: https://matplotlib.org/
- gpustats: https://github.com/wookayin/gpustat
The code uses VGG-19 preloaded neural net which can be downloaded from here: http://www.vlfeat.org/matconvnet/models/beta16/imagenet-vgg-verydeep-19.mat
python3 Main.py
Note: The project was developed under Jetbrains PyCharm IDE, in order to execute from here you have to run the Main.py.
Note2: The optical flow (deepmatching, deepflow) are not supported under Windows. It is recommended to use a Linux based operating system instead.
This project is part of my master's thesis. The documentation is written in hungarian using Latex. It can be built from doc\documentation\ folder with the following command:
pdflatex DeepArt.tex