Couldn't get it to find the unicode snowman[1]: <a href="http://i.imgur.com/gaIY9Gd.png" rel="nofollow">http://i.imgur.com/gaIY9Gd.png</a> (my drawing skills are awesome, no?)<p>But that aside, this looks like a neat idea. Not something I have any immediate use for myself, but could certainly be useful in some situations.<p>[1] <a href="http://www.fileformat.info/info/unicode/char/2603/index.htm" rel="nofollow">http://www.fileformat.info/info/unicode/char/2603/index.htm</a>
For the curious, this explains a lot of the science in a hands-on, approachable way: <a href="http://stackoverflow.com/questions/10168686/algorithm-improvement-for-coca-cola-can-shape-recognition" rel="nofollow">http://stackoverflow.com/questions/10168686/algorithm-improv...</a><p>EDIT: As this is a deep topic, there are also books if that's more your style:
<a href="http://www.amazon.com/dp/0123725380/?tag=stackoverfl08-20" rel="nofollow">http://www.amazon.com/dp/0123725380/?tag=stackoverfl08-20</a><p>or maybe you like Wikipedia (Ol' Trusty): <a href="http://en.wikipedia.org/wiki/Feature_detection_(computer_vision)" rel="nofollow">http://en.wikipedia.org/wiki/Feature_detection_(computer_vis...</a>
Damn that looks good, it worked for me for ② and u with umlauts.<p>Get the Japanese support in there and it will be amazing. What about using MS Mincho or MS Gothic for that? (It is free as in beer, but is the licensing off?)
My simple, clean ampersand (&) became a paperclip (0x1f4ce), a "fried shrimp" (0x1f364), several species of geometric triangle, and dozens of other silly, silly glyphs.<p><a href="http://i.imgur.com/NLkl75J.png" rel="nofollow">http://i.imgur.com/NLkl75J.png</a><p>the unicode block containing "fried shrimp" 0x1f364 -- why does this exist???<p><a href="http://www.unicode.org/charts/PDF/U1F300.pdf" rel="nofollow">http://www.unicode.org/charts/PDF/U1F300.pdf</a><p>Latest <i>Abstruse Goose</i> comic sums up my emotional response:<p><a href="http://abstrusegoose.com/496" rel="nofollow">http://abstrusegoose.com/496</a>
Very nice and amazingly accurate. But am I using your website the right way? When I look through the characters in the search results, if I see one that isn't rendering on my computer and just has a block instead of the symbol. I've been clicking on "bad" for "rate this suggestion". Thinking that it tallies up the total good/bad for a character to mean "how likely people are to have this character installed and working on their computers".<p>However, I now have a feeling that's not what that feature is for.
Reminds me of the excellent Detexify, which serves a similar purpose but for LaTeX symbols: <a href="http://detexify.kirelabs.org/classify.html" rel="nofollow">http://detexify.kirelabs.org/classify.html</a>
Tried to draw an alpha but it didn't get it.<p>So far I haven't tried Shapecatcher a lot but I think that <a href="http://detexify.kirelabs.org/classify.html" rel="nofollow">http://detexify.kirelabs.org/classify.html</a> works much better. Detexify is of course only for LaTeX symbols and doesn't do unicode.<p>To be useful Shapecatcher needs to become better at recognition.
<i>"If you can't find Chinese, Japanese or Korean glyphs, it is because I have yet to find a good free CJK font to use."</i><p>Are there not some CJK (or otherwise) fonts from, for example, Linux distributions that could have been used?<p>Or perhaps the emphasis could be on clarifying what is meant by "good" that deserves excluding such a large and useful character space for this type of application?
Sounds cool but sadly hasn't worked for the letters I often need but haven't easy access to (czech characters like: ř, ď and š). Perhaps because the element that makes them distinct from the latin (the "haček") is so tiny.<p>Edit, I just drew the ř larger and it recognised it correctly. Cool :)
Trying to draw U+1F4A9 (Pile of Poo). After several attempts, no luck.<p>I have learnt that Unicode contains even more weirdness than I thought before though, including 'Alchemical symbol for borax-3' (U+1f744), and 'doughnut' (U+1f369).
Nice idea, too bad that I tried to draw several variations of PI, and it showed me several interesting characters, but never a PI.<p>Seriously, it even showed some very PI-like things, but not PI itself. This is a downer.
Idea: instead of matching the shape of what the user has drawn raster-wise, let the user draw an svg-like path, and try to identify the letter by the trace.
Related:<p>A version of this for kanji that is very accurate<p><a href="http://kanji.sljfaq.org/draw-canvas.html" rel="nofollow">http://kanji.sljfaq.org/draw-canvas.html</a>
Does anyone know if the mirror image of this character: “ (U+201C) exists? I'm looking for a character that is the mirror image of the left double quotation mark, where the base is on the bottom and the character tapers from bottom right to top left. I don't know if any languages use that character.
Interesting thing, it recognized white queen correctly with diamonds on crown, but as black queen without diamonds and failed with only one diamond:<p><a href="http://i45.tinypic.com/3535j4x.png" rel="nofollow">http://i45.tinypic.com/3535j4x.png</a> — screenshot of variant with three diamonds.
Pretty neat, but I'm not sure what else it wants me to draw here <a href="http://i.imgur.com/rfU10rj.png" rel="nofollow">http://i.imgur.com/rfU10rj.png</a><p>Also it seems to fail on badly-drawn birds <a href="http://i.imgur.com/IZgrRkq.png" rel="nofollow">http://i.imgur.com/IZgrRkq.png</a>
Sadly it didn't recognize my clumsy attempt to draw the Look of Disapproval: <a href="http://i.imgur.com/0lvPFaJ.png" rel="nofollow">http://i.imgur.com/0lvPFaJ.png</a><p>On the upside, I learned that there is a Panda Face unicode character.
Really cool stuff. Luckily, I draw my lowercase 'a's in double-story form, but if you attempt to draw an a (or an accented a) with just a circle and a line, it's recognized as an 'o'.
Nice. It's not very good at finding faces though. The right face is there, but way down.<p><a href="http://i.imgur.com/PD1DkV6.png" rel="nofollow">http://i.imgur.com/PD1DkV6.png</a>
Really cool. The drawing tool would be even better if it drew as soon as you clicked (i.e., would draw a single dot if you click and don't drag the mouse).
It didn't recognize the capital letter A, only variations on the letter A and other unknown characters. I'm sure that the letter A is an unicode character!
Doesn't seem to include the drawing stroke order or anything, just the ultimate image. I guess if you know the strokes, you know the letter already, hm.
It claims Chinese/Japanese/Korean is unsupported, but works fine for me:<p><a href="http://i.imgur.com/AQBybDT.png" rel="nofollow">http://i.imgur.com/AQBybDT.png</a>
2 previous discussions.<p><a href="http://www.hnsearch.com/search#request/submissions&q=Shapecatcher&start=0" rel="nofollow">http://www.hnsearch.com/search#request/submissions&q=Sha...</a><p>Maybe the author can write a post detector next.