"Converts a raster image to a SVG image by converting blocks of pixels to
an SVG circle with a radius based upon the average pixel values of the
From: Matt Printz <hipok...@gmail.com>
Date: Thu, Mar 7, 2013 at 8:47 PM
Subject: [atxhs-discuss] Derasterizer, a raster image to SVG converter
At Tuesday's open house a few of you saw my latest project, a raster image
to SVG conversion app I wrote to make designs to etch with the laser
cutter. I got a lot of questions about it, so I decided to send it out to
At what point is it more efficient to vector cut a qr code, rather
than the actual text you want to communicate? Does anyone know how
long it takes to vector cut out some text, versus a qr code? I haven't
done any of the math, but it occurs to me that there might be a point
at which it would take longer to cut all of the circles than it would
to cut out the shapes for letters in a string of text in the first
place (just based on the entropy of qr codes, for example).
Turns out there's a (probably) more efficient ways to laser cut qr
codes using a much simpler technique: "As Fred Trotter has pointed
out, QR codes contain stencil islands in unpredictable configurations.
QR_STENCILER automatically detects and bridges these islands, using
thin lines that are minimally disruptive to the highly robust QR
algorithm. It does so through the use of two basic image processing
techniques: connected component labeling (sometimes called blob
detection) and 8-connected chain coding (sometimes called contour
On Thu, Mar 7, 2013 at 10:04 PM, Dustin Silverman <defende...@gmail.com> wrote:
> Well you could represent UTF8 as 8 dots arranged in 4 rows and 2
> columns, I'd say that would always be more efficient than actually
> cutting out letters.
Consider the following payload sans quotes: "testing". The number of
strokes is 9 or 10 depending on what font you are using.
When you generate a qr code and run it through derasterizer, it's much larger.