Experimenting with Image Recognition

06/30/15

So, as I've mentioned, I was on vacation the past week. About the only thing I didn't take a break from was collecting my receipts. I did, however, neglect to put them into my JSON file, and now I have quite the pile to log by hand.

I believe now is the time for Enlightened Laziness. Though, I'll have to also do it by hand while I implement such laziness.

In any case, I'd like to see how easily I could go about translating one of my receipts into parseable text. Ideally, I could do this with my phone, and send that text elsewhere when my phone is connected to the Internet.

I'm going to want to figure out the best way to capture a receipt with my phone. It may be possible to make a multipage pdf with some sort of video-to-image thing, or using that panorama algorithm. That should be something that is already available, I hope.

Anyways, what I'll try for now is to do 10 minutes of image recognition research for every ten minutes of inputting the receipt by hand. Let's see how fast I can automate my tedious work.