The new restaurant poses an ordeal for me. Trudging towards the cashier, I pretend to read the menu behind her; image, however blurred, is everything. I rack my brain for memories. Surely somebody suggested something. Sadly, whatever recommendation whoever gave remains lost in a fog. I hopefully search the nearby wall for a poster. A sign loudly proclaiming some dish, any dish I can order and claim to have actually wanted all along. My expression, blank as the walls around me, remains fixed as I begin thinking of probability, the odds of picking a random item and guessing right. Too late to think of one now; it’s the end of the line and the cashier waves at me. I walk forward, making what hopefully looks like eye contact, gesture towards the menu, and say ‘Sorry, my vision isn’t great, is there something closer?’
	An accurate, if softened description of my vision. I am legally blind, but blindness is not binary; it comes in many shades. Perception does not follow legality. In the eyes of medicine and law I am blind, yet those who know me know I can see. I excel at fooling people. Wearing earbuds and feigning obliviousness to imply that I could have seen you in the crowd if I was paying attention. Pretending to text while using my phone to magnify something too small to read. Small tricks to conceal reality. This is what drove me to study computer science.
	After nearly a decade of fruitlessly squinting at menus behind befuddled cashiers, I stopped trying to deny my blindness and started trying to fix it. A plethora of overpriced devices exist to help the under-visioned. From a simple magnifying glass to an advanced CCTV, these devices do the same thing: magnify a target. But when I’m trying to read a chalkboard magnification is almost worthless. The field of view is so reduced that following anything becomes an ordeal. What I need is the text. With a digital copy of the text, I can format it however I like, but most professors are not going to give out a copy of their lecture notes. Rather than rely on notetakers, I started self-studying computer vision in order to make a program to digitize text from an image. Using Google’s open source Tesseract optical character recognition API I was able to write an android application, but handwriting was very difficult to consistently identify.