artificial-artwork 2.0.0

Create artificial artwork by transfering the appearance of one image (eg a famous painting) to another user-supplied image (eg your favourite photograph).

Neural Style Transfer

Create artificial artwork by transferring the appearance of one image (ie a famous painting) to another user-supplied image (e.g. your favorite photo).

Documentation: https://boromir674.github.io/neural-style-transfer/

+ =

Uses a Neural Style Transfer algorithm to transfer the style (aka appearance), from one image, into the contents of another.

Neural Style Transfer (NST) is an algorithm that applies the style of an image to the contents of another and produces a generated image.
The idea is to find out how someone, with the painting style shown in one image, would depict the contents shown in another image.

NST takes as INPUT a Content image (e.g. a picture taken with your camera) and a Style image (e.g. a Van Gogh painting) and Generates a new image.

Overview

This project provides an NST algorithm through

• the artificial_artwork Python package (aka module)
• the nst CLI
• the boromir674/neural-style-transfer Docker image

Build	Package	Containerization	Code Quality

Features

• VGG-19 Convolutional Neural Network, as Image model
• Selection of style layers at runtime
• Efficient Iterative Learning Algorithm, with tensorflow
• Fast minimization of loss/cost function with parallel/multicore execution
• Selection of Algorithm Termination Condition/Criteria, at runtime
• Periodic persistance of Generated image, during Learning loop

Quick-start

Run a demo NST, on sample Content and Style Images:

mkdir art
export NST_HOST_MOUNT="$PWD/art"

docker-compose up

# Process runs, in containerized environment, and exits.

Check out your Generated Image! Artificial Artwork: art/canoe_water_w300-h225.jpg+blue-red_w300-h225.jpg-100.png

xdg-open art/canoe_water_w300-h225.jpg+blue-red_w300-h225.jpg-100.png

Usage

Run the nst CLI with the --help option to see the available options.

docker run boromir674/neural-style-transfer:1.0.2 --help

Development

Installation

Install nst CLI and artificial_artwork python package from pypi:

Note: Installation on Debian-based Distros for Python 3.11 require distutils which is not included in python3.11 standard distribution (but included in python3.10).

sudo apt-get install python3.11-distutils

pip install artificial_artwork

Only python3.8 wheel is included atm.

Sample commands to install the NST CLI from source, using a terminal:

git clone https://github.com/boromir674/neural-style-transfer.git
    
pip install ./neural-style-transfer

The Neural Style Transfer - CLI heavely depends on Tensorflow (tf) and therefore it is crucial that tf is installed correctly in your Python environment.