A new algorithm for retrieving topological skeleton as a set of polylines from binary images: WASM version

Usage no npm install needed!

<script type="module">
  import skeletonTracingWasm from '';


Skeleton Tracing WASM

by Lingdong Huang


<script src=""></script>
const TraceSkeleton = require('skeleton-tracing-wasm')

import TraceSkeleton  from 'skeleton-tracing-wasm'

You first need to initiate/load the wasm module calling a static method load()

const tracer = await TraceSkeleton.load()

then you can use the API methods

const { polylines,rects } = tracer.fromCanvas(HTMLCANVAS)

Or instead of async/await you can do

TraceSkeleton.load().then(tracer => {
    // wasm module loaded
    // here is your code 
    const { polylines,rects } = tracer.fromCanvas(HTMLCANVAS)


The below API's take an image representation and returns an object holding the polylines as well as rects processed by the algorithm (the latter is mainly for visualization)

    "polylines": [[[x,y],[x,y]],[[x,y],[x,y],[x,y],...],...],
    "rects":     [[x,y,w,h],[x,y,w,h],...]


Takes in an HTML Canvas object and returns the skeleton as polyilnes as well as the rects.


Takes JavaScript ImageData object (e.g. document.createElement("canvas").getContext('2d').getImageData(0,0,100,100))


Takes array of booleans (or truthy and falsy values), e.g. [0,1,0,1,1,1,0,0,...] or [0,255,255,0,...] or [true,false,true,false,...] or even [undefined, "ok", null, "yes", ...]


Takes in a (char*) such as "\0\1\0\0\1\1\0....". This is the fastest (though probably most obscure) API because it does not need to translate the input to C constructs.

TraceSkeleton.visualize(result, {scale, strokeWidth, rects, keypoints})

Conveniently visualize the result from the previous functions, returns a string holding an SVG (scalable vector graphics).


  • scale: factor to scale the drawing
  • rects: draw the rects?
  • keypoints: draw the keypoints on the polylines?
  • strokeWidth: weight of the polyline strokes.

See /index.html for more detailed usage example, with animation, interactivity, webcam, etc.

More info at

Github Original Project

Developed at Frank-Ratchye STUDIO for Creative Inquiry at Carnegie Mellon University.