While that's true, having the ability to capture strokes now allows machine-learning models to better determine what potential strokes were used to make a specific shape. Just because we didn't have it for everything doesn't mean it's not useful for adding accuracy to the past.
The digital medium often strips human expression when capturing input.
Devices such as the Apple Pencil are getting better at capturing pressure and highly-detailed paths, which inturn are encoded as vectors and can gather information that is lost in physical mediums, such as coordinates, drawing order, and temporality, even capturing things not present on paper such as hover gestures and pressure.
Yet it's common practice to reduce gestures and clicks to points, lines, or curves - two clicks from thousands of users in a CAD environment may output the exact same line from point A to point B - which often forget about the expressiveness of pencil strokes on a sheet of paper, features which could be used by machine learning algorithms to discern intent in the input of different users.
I was too loose with that. There is CLIPDraw and others that operate at the stroke/action level but haven't been trained on as much data. Still impressive at the time:
> My dream drawing app would let you draw a stroke multiple times and then pick out your best attempt. No need to undo all the bad ones to try again. It separates the generation and the judgment.
Just use SVG-edit & draw each stoke on separate "Layer".
Before there was a dense scientific paper. Now they've released an incredibly simple tool that lets anyone draw photorealistic images with simple strokes.
Saying that is repackaging something into an easier-to-use tool seems like quite a stretch. They didn't put a GUI on curl or something.
Hello,
I'm Kyle, founder of drawing navigator (www.drawnav.com), design engineer, programmer, and maybe even something interesting one day.
After working as an engineer for almost a decade I picked up on a persistent problem while working with old blueprints. Old drawings were incredibly difficult to find, sometimes impossible, even if they were scanned in.
I tend to use existing data to build up the underlying drawing hierarchy and be able to find whatever I'm looking for without hunting for hours. Slowly patching together the functionality to automatically generate the database from scanned prints.
I have been using this software alone for quite a while before I decided to see if anyone else thought it was useful in the same way that I do.
Here to discuss any questions or comments anyone has.
right, but for a square, you have to add 8 samples, not 2, to handle the 4 starting points and 2 directions, but this does not account for the users who multi-stroke
> Different strokes...
I see what you did there :] I'm definitely in the reduce user burden camp.
Sad reality is, people expect more. There's something satisfying about being able to click a 'draw shape' icon on a tool bar and then click and drag to draw the shape on a canvas. That's an expectation that dates back to at least MacPaint in 1984, and it's a huge difference from drawing a shape using a rendering plugin buried under three menu levels.
My dream drawing app would let you draw a stroke multiple times and then pick out your best attempt. No need to undo all the bad ones to try again. It separates the generation and the judgment.
I wonder why they didn't take advantage of the fact that it could draw a huge amount of lines in a small amount of time and tweak their edge-detection algorithm to draw more lines.
Anyway, I'm sure there was very good reason why they haven't... likely the lines they can draw can't be very fine with the sort of marker they have to use.
One of the things I want to do is to use autotrace centerline to get a basic stroke trace, and then manually tweak it so that drawing the glyphs makes sense. Then I can use that for SVG animations (there are Javascript tutorials for that), which I can use to help my kiddo learn how to write. (Whether I should teach her to forge my handwriting so early might be another matter... ;) )
But that's an implementation detail - some software might let you hand write text into an image with a brush or pen, and that might use as its internal representation the X,Y coordinates of the brush strokes.
A lot of great comments here on the content and presentation.
I'm really interested, also, in the technical aspects of how this was put together.
For instance, the app was able to recognize the shape of the line I drew. On my first attempt, it correctly identified that my line was an S shape; then when I started fiddling around, it was able to identify that my line was (roughly) straight. I wonder what method they used to identify those shapes.
reply