image2braille - convert images to braille

Super simple command line program that converts an image to braille and outputs it to the terminal or a file.

Just started learning Go so i was trying to find something interesting to do. The code is pretty straight forward and it was fun to code.

Table of contents

• Usage
• How it works
• Resources

---

Usage

Assuming you have Go installed (https://golang.org/doc/install) go ahead and compile the code:

go build

Run it with --help to see all the available options:

./image2braille --help

Usage of ./image2braille:
  -image string
    	path to image file (*required)
  -threshold int
    	the color (0-255) threshold to convert a pixel to a dot (default 128)

Just fyi you do need to play around with the threshold to figure out what works and what doesn't for some images. I found that although most recommend 128 which i did set as a default higher values such as 150-199 worked better.

---

How it works

The Braile Unicode codepoints (0x2800-0x28FF)

Using the extended set there is a character for a every possible combination 8 dots in a 2x4 grid.

Braile Patterns:

Braille Pattern        Hexadecimal value of each dot
    #######                     #########
    # 1 4 #                     # 1   8 #
    # 2 5 #                     # 2  10 #
    # 3 6 #                     # 4  20 #
    # 7 8 #                     # 40 80 #
    #######                     #########

Originally Braille characters had 6 dots. The two lowest dots were added later that's why the numbering is off for dot 7 and dot 8.

Enconding to a bit string is simple enough:

#######
# * * #
# * * #
# - - #
# * - #
#######
1 = *, 4 = *, 2 = *, 5 = *, 3 = -, 6 = -, 7 = *, 8 = -

Braille Pattern:        **-**-*-
Dot Numbers available:  12345678
Dots Raised:            11011010
=================================
Binary Value:			01011011 	(by reversing the order of the dots raised)
Decimal (base 10): 		91
Hex (base 16):			0x5b
Braille Character:		0x5b + 0x2800 = 0x285B = '⡛'

We can compute the mapping by adding together the hexadecimal values for the dots raised. Either computing directly in hexadecimals or in binary the result is then added to 0x2800 the offset for the Braille Patterns Unicode block.

In the code i used a pixel map which is just a two dimensional int array ([4][2]int) to store the mapping for the hexadecimal values.

Going through the image we grab each pixel and find a grayscale value for it using some of Go's built-in functions. Depending on the threshold we then decide whether a dot is raised or not.

That's basically it :)

Anyway hope you enjoy using it.

---

Resources

• Braille Patterns - https://en.wikipedia.org/wiki/Braille_Patterns
• Braille Unicode Pixelation - https://blog.jverkamp.com/2014/05/30/braille-unicode-pixelation/