A few weeks back while being totally fine doing nothing on the couch I had been handed over a pile of stickers by my wife and instructed to help her find missing stickers to fill up a sticker book for our kid.
Her method was to memorize one number and iterate hundred of stickers which I discarded immediately and solemnly announced v2 by memorizing 2 numbers! v2 lasted one iteration then v3 was announced, lasted 10sec and then ragequit. 🤯
Classic scratch your own itch moment, I've decided to solve that problem with an app:
- input field to enter the numbers of stickers that are not yet in the book
- camera view that detects and notifies when number is matched
As I've already used Vision framework in one project for image classification this was a great opportunity to try out Vision's text detection. As a head start I've leveraged official sample code Reading Phone Numbers in Real Time.
Initial goal was to build the app with number recognition for those stickers but than I realized that after some adjustments I could use it as a card scanner in my Beleep app so I separated the camera and recognition logic to framework and called it weezee, currently it supports numbers only.
The main goal is to have view controller similar to UIImagePickerController with ability to recognize three data types: text, numbers and bar codes.
Here's demo video of weezee framework justifying it's purpose:
