text to Braille is a whole other problem. Sure, you can convert the letters to Braille dot-forms, but real Braille uses a whole load of contractions and conventions to make the text quicker to read.
The ENAbling MEdia for braille Literacy (ENAMEL)
project at York University has a low-cost Braille display that they've managed to interface with Raspberry Pi. It was barely past concept stage last time I saw it, but the individual elements were cheap, even if they needed a voltage conversion in the 50-200 V range (well into the ow!
range if you touched the wrong terminal) and worked decently. An yes, it used solenoids, but in tiny Braille display elements made in Germany.
Though I work
in assistive technology, Braille has a whole level of history, culture and complexity that I know I'm no expert in. I'm currently investigating tactile printing for vision loss (trying to use a 3D printer doesn't seem to be ideal) to help a friend who is a musician gain access to his scores, but he doesn't read Braille at all.
HermannSW wrote: ↑
Tue May 22, 2018 9:35 pm
Code: Select all
$ jpegtopnm test.jpg | pnmflip -rotate90 | pnmtojpeg > bof.p9.jpg
a quiet voice suggests jpegtran
to make this much easier. Tesseract has ways of detecting and correcting rotation, but they're not infallible. Tesseract also works slightly better with greyscale rather than one-bit images, contrary to older received wisdom.
‘Remember the Golden Rule of Selling: “Do not resort to violence.”’ — McGlashan.