Braille#

This style uses the characters of the Braille writing system.

Example

Consider the following image:

Apple Computer [Rob Janoff, 1977]

Here's what it should look like:

Procedure#

This style is implemented using the BrailleDrawer.

Initialization#

Threshold#

The threshold parameter filters out pixels of the input image whose grayscale intensities are lesser than it.

Consider the following image:

Here's what it should look like:

threshold = 0

70

108

168

210

Matrices#

The kernel attribute holds a NumPy ndarray containing the following matrix:

The Unicode encoding of the 8-dot cell Braille system is done by assigning each of the dots a power of 2. Each character in the Braille Patterns block has a unique Unicode value that is obtained by summing these powers.

The charset_array attribute holds another NumPy ndarray containing all 256 Braille characters.

Conversion#

Resizing#

Assuming the output text should have the dimensions text_height and text_width, the image must be resized according to the following criteria:

• image_height = 4 * text_height.
• image_width = 2 * text_width.
• If either image_height or image_width is 0, it is derived from the other by preserving the aspect ratio of the original image.

Following the above algorithm, each pixel of the resized image will be assigned to one dot (Braille character dot) in the output text.

Processing#

1. The resized image is first converted to its grayscale.
2. Each pixel is set to either 0 or 1 based on whether its grayscale intensity is below or above the threshold.
3. A convolution operation is performed on this filtered image using the kernel matrix. The resultant matrix has the ofsetted Unicode values for the corresponding Braille character.
4. The charset_array is indexed with the resultant "indices" matrix, giving the final text_matrix.