Created for Tartanhacks 2020
Built using Python, Pytorch, OpenCV, pytesseract modules and CRAFT-pytorch.
People with dyslexia sometimes aren't able to read important text in public places, such as safety instructions, lecture notes, and street signs.
This application simply takes a picture provided by the user, detects regions with text, extracts the text, and re-displays the image with captions written in dyslexic-accessible font.
Easy as pie!
Challenges Getting the dyslexic font to work was surprisingly difficult. Also I had originally intended to do this on top of a live video stream but the computation was quite a lot.