About the process
Since the first day of the publication I have been playing with Google’s Quick, Draw! dataset and can’t get enough of it. My brave laptop spent nights and nights computing letters and scenes from random subsets of doodles (way over 300.000 in sum by now). The idea behind the algorithm is pretty simple: I find an empty space for a doodle by going through all the pixels linearly, scale it as big such that it still fits inside the area and continue with the next random doodle. Here are a few results, short videos I made to illustrate the algorithm and some scenes that you can download in poster format.
Screencasts / How-To
I arranged each picture on A1 for easy printing, I did not test it yet but the size should fit well.
Please take a picture if you hang this up somewhere or do something with it and send it to me or post it in the comments!
License (downloadable posters only)
This work by Deborah Schmidt is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Based on a work at https://github.com/googlecreativelab/quickdraw-dataset.