Convert images to beautifully crafted ASCII art using the power of neural networks. All the fancy stuff / hard work is done in ascii-net by @wahtak.
Turn an image
into ASCII
@%$,
#%%F
___________ _,
,,wggW# ,@%%%%%%%%%%%%%%%%%#
%%% $M***3******%%%*^^^`
.%%Y * $ g%%$
$%% aF $%%F
,%%F ,# g%%$
%%$ %F %%%F
{%%F $7 +%%$ g
%%# ,M aL _g%7 $%%) aC
(%%%w#` %%%%%* V%%L_,e`
______ **^ `**^ `***`
_,,g@%%###%%%%%y_
_,s%%#*` `%%%%%k
,@%%%#` %%%%%L
w%%%%* ,%%%%%
,%%%%#` %%%%%#
g%%%%# ,%%%%%7
,$%%%%Y ,@%%%%#`
,%%%%%# ,g%%%%#*`
%%%%%% _,,s%%%##*`
$%%%%%c,,,gg@%%%##*^
y%%%%%%MM**^^`
$%%%%%%
%%%%%%%
%%%%%%% ,
V%%%%%%% ,g%7`
%%%%%%%%,_ ,g$%*
%%%%%%%%%g,,____,,ws%%#^
*#%%%%%%%%%%%%%%%#M^
`*M##%%%%###*^
For some extra fun, try piping the result through lolcat:
git pull <this repo>
cd img2ascii
docker build -t img2ascii:latest .
Put all the images you want to asciify in a folder at /path/to/images (Protip: use only one image or you'll have a hard time finding results other than the last one. Output is pretty messy).
Then run
docker run --rm -t --env THRESHOLD=50 -v /path/to/images:/images img2ascii
Change the threshold for the black/white conversion via the $THRESHOLD environment variable.
- option to change output sizes
- variable number of optimization runs
- can i haz colors? <-- WARNING: This might result in some real work.
- It's 2017, so give me a goddamn Webinterface. And a REST-API
- ok but don't you think we should start with writing the result to a file instead of stdout?
I read Turing's 1936 paper on computing and a page of Javascript example code and guessed at everything in between. If this code breaks anything or annoys your cat, don't blame me. Please :)
This repo is just my 2 lines of glue code. All credit goes to ascii-net by @wahtak