Skip to content

Latest commit

 

History

History
53 lines (33 loc) · 2.33 KB

README.md

File metadata and controls

53 lines (33 loc) · 2.33 KB

tinqta

AI created illustrations in your own scribble style!

  • easy and quick: train models on your own style
  • transfer images into your scribble style
  • creates beautiful vector graphics: small in size and easy to adjust or animate

🧑‍💻 Installation

docker (recommended)

Simply clone the repository and install via docker compose up . Afterwards call localhost:8000 from your browser.

local install with python env

Clone the repository andstart a python environment. Then run to install: pip install -r requirements.txt

Afterwards you can start tinqta with python server.py

✏️ Usage

First, choose either a photo from you PC or take a snapshot via webcam. In a next, automatic step, tinqta will remove the background of your image.

Adjust edge detection parameters to your taste. Less is often more :) Also choose a model that generates shapes for shadow areas in your image.

edge.mp4

In a last step you can adjust colors and style of your generated image. Edges and shadow pattern lines can be adjusted separatly. Also you can set an animation length. Either just save your image as an .svg now or click the copy icon to save the html code to your clipboard, so you can directly paste it for example into a website.

fin.mp4

(optional) In the Model selection, you can also opt to train a scribble model yourself. Just draw shapes you like on the canvas, choose a name and hit "train": in the next few minutes you can enjoy the sight of neurons fitting your lines ;)

train.mp4

🎨 Examples

examples

👀 Projects using tinqta

❤️‍ Acknoledgements

Media Lab KI Salon