yuqing-liu-dut / live-pytorch

Towards Layer-wise Image Vectorization

LIVE-pytorch

Towards Layer-wise Image Vectorization

Updated for rebuttal (Jan/28/2022)：

User study

We create a user study as suggested. A more complex user study will be added in the revised version.

The results are collected here: user study details

Code installation

we added detailed conda env file and collected detail system information to help the installation.

A more detailed docker and Google Colab demo will be provided.

LIVE is able to explicitly presents a Layer-wise representation for simple images.

Installation

pip3 install torch torchvision
pip install svgwrite
pip install svgpathtools
pip install cssutils
pip install numba
pip install torch-tools
pip install visdom
pip install scikit-fmm

Next, please refer DiffVG to install pydiffvg

Run

python main.py --config config/all.yaml --experiment experiment_5x1 --signature demo1 --target data/demo1.png

Please modify the config files to change configurations.