Home

Awesome

General Virtual Sketching Framework for Vector Line Art - SIGGRAPH 2021

[Paper] | [Project Page] | [中文Readme] | [中文论文介绍]

This code is used for line drawing vectorization, rough sketch simplification and photograph to vector line drawing.

<img src='docs/figures/muten.png' height=300><img src='docs/figures/muten-black-full-simplest.gif' height=300>

<img src='docs/figures/rocket.png' height=150><img src='docs/figures/rocket-blue-simplest.gif' height=150>    <img src='docs/figures/1390.png' height=150><img src='docs/figures/face-blue-1390-simplest.gif' height=150>

Outline

Dependencies

Testing with Trained Weights

Model Preparation

Download the models here:

Then, place them in this file structure:

outputs/
    snapshot/
        pretrain_clean_line_drawings/
        pretrain_rough_sketches/
        pretrain_faces/

Usage

Choose the image in the sample_inputs/ directory, and run one of the following commands for each task. The results will be under outputs/sampling/.

python3 test_vectorization.py --input muten.png

python3 test_rough_sketch_simplification.py --input rocket.png

python3 test_photograph_to_line.py --input 1390.png

Note!!! Our approach starts drawing from a randomly selected initial position, so it outputs different results in every testing trial (some might be fine and some might not be good enough). It is recommended to do several trials to select the visually best result. The number of outputs can be defined by the --sample argument:

python3 test_vectorization.py --input muten.png --sample 10

python3 test_rough_sketch_simplification.py --input rocket.png --sample 10

python3 test_photograph_to_line.py --input 1390.png --sample 10

Reproducing Paper Figures: our results (download from here) are selected by doing a certain number of trials. Apparently, it is required to use the same initial drawing positions to reproduce our results.

Additional Tools

a) Visualization

Our vector output is stored in a npz package. Run the following command to obtain the rendered output and the drawing order. Results will be under the same directory of the npz file.

python3 tools/visualize_drawing.py --file path/to/the/result.npz 

b) GIF Making

To see the dynamic drawing procedure, run the following command to obtain the gif. Result will be under the same directory of the npz file.

python3 tools/gif_making.py --file path/to/the/result.npz 

c) Conversion to SVG

Our vector output in a npz package is stored as Eq.(1) in the main paper. Run the following command to convert it to the svg format. Result will be under the same directory of the npz file.

python3 tools/svg_conversion.py --file path/to/the/result.npz 

Important Notes

In SVG format, all the segments on a path share the same stroke-width. While in our stroke design, strokes on a common curve have different widths. Inside a stroke (a single segment), the thickness also changes linearly from an endpoint to another. Therefore, neither of the two conversion methods above generate visually the same results as the ones in our paper. (Please mention this issue in your paper if you do qualitative comparisons with our results in SVG format.)

<br>

Training

Preparations

Download the models here:

Download the datasets here:

Then, place them in this file structure:

datasets/
    QuickDraw-clean/
    QuickDraw-rough/
    CelebAMask-faces/
outputs/
    snapshot/
        pretrain_neural_renderer/
        pretrain_perceptual_model/

Running

It is recommended to train with multi-GPU. We train each task with 2 GPUs (each with 11 GB).

python3 train_vectorization.py

python3 train_rough_photograph.py --data rough

python3 train_rough_photograph.py --data face
<br>

Citation

If you use the code and models please cite:

@article{mo2021virtualsketching,
  title   = {General Virtual Sketching Framework for Vector Line Art},
  author  = {Mo, Haoran and Simo-Serra, Edgar and Gao, Chengying and Zou, Changqing and Wang, Ruomei},
  journal = {ACM Transactions on Graphics (Proceedings of ACM SIGGRAPH 2021)},
  year    = {2021},
  volume  = {40},
  number  = {4},
  pages   = {51:1--51:14}
}
<br>

Projects Using this Model/Method

Painterly style transfer (TVCG 2023)Robot calligraphy (ICRA 2024)
<img src="docs/figures/applications/Painterly-Style-Transfer.png" style="height: 170px"><img src="docs/figures/applications/robot-calligraphy.png" style="height: 170px">
Geometrized cartoon line inbetweening (ICCV 2023)Stroke correspondence and inbetweening (TOG 2024)
<img src="docs/figures/applications/Geometrized-Cartoon-Line-Inbetweening.png" style="height: 170px"><img src="docs/figures/applications/Vector-Line-Inbetweening2.png" style="height: 170px"><img src="docs/figures/applications/Vector-Line-Inbetweening-dynamic1.gif" style="height: 170px">
Modelling complex vector drawings (ICLR 2024)Sketch-to-Image Generation (UIST 2024)
<img src="docs/figures/applications/complex-vector-drawings.png" style="height: 170px"><img src="docs/figures/applications/sketch-to-image.png" style="height: 170px">