• About DIVISIO
  • AI Basics
  • Amazon DJL - a new DL framework for Java
  • BGL symposium 2019 - lecture 'AI and Magic'
  • Categories
  • ChatGPT
  • ChatGPT for Teams: Privacy-Compliant Use in the Workplace
  • Conferences
  • Contact
  • Deep Fakes - How to spot faked Images
  • Deep Java Learning Introduction - Part 1: NDManager & NDArray
  • Depixelizer
  • DIVISIO Blog | Understanding AI, DL and ML together
  • DJL
  • DL4J Workshop at the ML Summit in Berlin
  • Enterprise Tensorflow - Java vs. Python
  • Enterprise TensorFlow 2 - Saving a trained model
  • Enterprise TensorFlow 3 - Loading a SavedModel in Java
  • Enterprise TensorFlow 4 - Executing a TensorFlow Session in Java
  • Enterprise Tensorflow: Code Examples
  • Git as a management tool for training data and experiments in ML
  • GPT-3
  • GPT-4
  • Homepage
  • Imprint
  • Java
  • Jax 2018 - Talks about DL4J and more
  • Jax 2019 Recap
  • KI Basics
  • Konferenzen
  • ML Conference 2017 in Berlin
  • MLOps: Establishment and operation of an AI
  • Neural networks - The five most common mistakes
  • NLP
  • NLP, NLU and NLG: AI and text
  • OpenAI API
  • Privacy Policy
  • Recap: ML Conference 2019 in Munich
  • Sokratext NER - Analyze German texts using Java
  • Sokratext NLG - Generate thousands of German texts with just one click
  • Sokratext POS Tagger - Automated Word Type Recognition with Java
  • Sokratexts
  • Tags
  • TensorFlow
  • TensorFlow and Java - An interview with entwickler.de
  • Types of Artificial Neural Networks
  • Understanding AI - Part 1: What is AI?
  • Understanding AI - Part 2: Symbolic AI, Neural Networks and Deep Learning
  • Understanding AI - Part 3: Methods of symbolic AI
  • Understanding AI - Part 4: The basics of Machine Learning
  • Understanding AI - Part 5: Supervised & Unsupervised Learning in ML
  • What are Neural Networks and how do they work?
  • Whisper 3 Large for JAVA
de en

DIVISIO Depixelizer for your retro pixel art

Depixelize your images and convert them into scalable vector graphics - with DIVISIO’s Depixelizer. Learn more about our tool, how it works and what led us to this idea.

A unique view at Pixel Art - Experience your retro computer game graphics in a new dimension.Our special computer vision algorithm is optimised for low-resolution and brightly coloured images. It is implemented as an online tool, so you can use directly in your browser. The result is a free scalable vector graphic (SVG).
If you want to know how exactly our AI-system works read more in our step-by-step summary.

How does the Depixelizer algorithm work?

The Depixelizer is our implementation of the paper “Depixelizing Pixel Art” by Johannes Kopf and Dani Lischinski. This algorithm converts “pixel art” into freely scalable SVG graphics by going through the following steps:

a) The pixels are loaded and converted into the HSV colour space. (https://de.wikipedia.org/wiki/HSV-Farbraum) b) Based on their colour values pixels are connected in a graph. [(https://de.wikipedia.org/wiki/Graph_(Graphentheorie)](https://de.wikipedia.org/wiki/Graph_(Graphentheorie) c) A heuristic determines connected areas in the graph, based on the “Gestalt laws”. (https://de.wikipedia.org/wiki/Gestaltpsychologie) d) The outlines of the surfaces are turned into splines. (https://de.wikipedia.org/wiki/Spline) e) The splines are smoothed using an energy minimisation approach and similar colours are converted into colour gradients. (https://en.wikipedia.org/wiki/Energy_minimization)

In order to remain SVG-compatible, we have dispensed with the creation of the colour gradients.

The idea behind the project

For us, software development and algorithms are creative, exciting and even artistic. Even though we primarily put our energy into our own products - such as the Sokratext engine for German language - and, of course, our clients' software, we sometimes come across a paper that is simply too interesting not to implement. And it is precisely with such implementations that we try out new technologies and techniques from which you as a DIVISIO customer benefit. Why not depixelize your favourite Nintendo character such as Super Mario, a character from Pokemon or Minecraft? We hope you enjoy playing with our depixelizer!

Data protection notice: All uploaded images and the results will be deleted immediately after depixelizing. No pictures will be saved on our servers.

IMPORTANT! This algorithm works well with high-contrast pixel art in original resolution - photos, JPEGs with compression artefacts and upscaled pixel images do not produce usable results.