Skip to content

A fast and reliable image to ASCII art converter that supports all major image types

License

Notifications You must be signed in to change notification settings

Uki99/img2ascii

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

img2ascii

img2ascii is ASCII art generator that accepts most raster image formats written using GDI+ and Windows API

  • JPG
  • PNG
  • GIF
  • BMP
  • TIFF

Features

It's command line based program, which means you supply options via console.

-w{number} specifies number of characters in a single line of ASCII art. Maximum is 400. You can use predefined templates: -tiny, -medium and -large instead.

-s{number} specifies at which width the image should be scaled preserving aspect ratio before processing. Smaller resolution will yield faster processing.

-g specifies that to the image is applied negative gamma correction and contrast is boosted. This could help flat, overly exposed or underexposed images have more defined edges!

-e (edge detection) specifies that the program should use edge detection and output improved art by using more complex algorithm for generation Processing time is slightly increased.

Alt text

Algorithms

The program features 2 distinct algorithms for art generation.

Before any processing the image is scaled up or down according to the input. Scaling the image down could result in much faster processing whereas scaling up could result in finer detail showing in the final art.

The default algorithm is faster to execute. It takes the scaled image and devides it into grid, where each section of that grid is a portion of the image. Grid size is calculated according to the ammount of characters per line in the art and image width in pixels. The section luminance (relative luminance) is then calculated and normalized, thus having quantitative value which is then compared to the character map where right character will be chosen to represent that section.

Edge detection algorithm is basically further division of already mentioned sections into 9 non equal parts with predetermined, constant ratios. Each part of that section will have its relative luminance calculated and normalized. From data structure point of perspective, it's just an array of 9 floating point numbers which represent normalized luminance of the section. That array is compared to different custom and precalculated character map which consists of each printable ASCII character represented by 9 floating point numbers that are generated much like section luminance is. The characters luminance map is precalculated and optimized using Terminal 8x12px monospaced raster font, which should be available on any Windows machine.

Comparing each section to the character map is basically searching for the most similar vector to the provided one out of the bunch. Best results after are given using Euclidean distance algorithm.

Alt text

TO DO:

  • Implement multithreading which should speed up the processing
  • Add colour to characters which would be averaged out from sections yielding coloured ASCII art

About

A fast and reliable image to ASCII art converter that supports all major image types

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published