lowpolify

“I’ve been waiting for you, Obi-Wan. We meet again, at last. The circle is now complete. When I left you, I was but the learner; now I am the master.” 😎😎😎

PS. Formal introductions are yet to begin, in case you were wondering. 😅

Introduction

The goal of lowpolify is quite simple. Generate low poly versions of any given image.

Now what is 'low poly' you may ask.

Here's what Wikipedia has for you: Low poly is a polygon mesh in 3D computer graphics that has a relatively small number of polygons. Low poly meshes occur in real-time applications (e.g. games) and contrast with high poly meshes in animated movies and special effects of the same era. The term low poly is used in both a techni... yada yada yada you get the idea. Basically objects composed of only polygons. 💁💁

If you'd like to skip all the chit chat, here's the DEMO. Arr, matey, thar be ye booty!
https://lowpolify.ghostwriternr.me/

Approach

Take an image, lowpolify it and poof!! 💥 LOW POLY, BABY!! 💃💃

Here's an ELI5 TLDR 👶 :

• Pre-process the input image to remove noise and reduce image size
• Detect edges in the input image
• If the image contains humans faces, detect facial features as well
• Choose a random subset of the above detected points / edges
• Triangulate using Delaunay Triangulation
• Fill the triangles with the mean value of all pixels contained by it (in parallel for faster computation)
• LOW POLY, BABY! 💃💃

Sample output

Lo and Behold!

High-poly input	Low-poly output

Screenshots

Now everybody loves to see the master at work, right? Right?? AMIRITE???

Dependencies

• Node.js
• Python3 (Along with the following python modules):
  • cv2
  • numpy
  • scipy
  • sharedmem
  • dlib

How to use

If you only need to run the script

• Download the core script from here.

wget https://raw.githubusercontent.com/ghostwriternr/lowpolify/master/scripts/lowpolify.py

• Download and extract the .dat file from dlib for facial features detection. Place the extracted file under scripts (the same directory as the lowpolify.py script).

• Run the python script as:

python lowpolify.py <input-image> <output-image> <cFraction>

Here, cFraction denotes the fraction of points from original edges to be taken for triangulation.

If you would like to use the webapp

• Clone the repo

git clone https://github.com/ghostwriternr/lowpolify

• Navigate into the cloned repository

cd lowpolify

• Install all npm modules

npm install

• Create a file secrets.json under the app/secrets/ directory. The structure of the file should look like this:

{
  "apiKey": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
}

You can create your own apiKey at your Clarifai Developer account page.

• Download and extract the .dat file from dlib for facial features detection. Place the extracted file under scripts (the same directory as the lowpolify.py script).

• Start the server using node

node server.js

• Marvel at the beauty served hot at http://localhost:8080/

License

MIT 🎓

Psst

In case you too got bored reading this README, foxy here wants to give you a hugsy!