Optical Character Recognition is used to allow computers to read like humans do. For example, if you took a photo of a car number plate you (as a human) can read each letter and number. However all a computer sees is a load of pixels. OCR is used to educate a computer to recognize a bunch of pixels is in fact a letter.
If you haven't already, I strongly recommend you watch the video to understand more about how the project works.
As this technology is still new I will continue to keep improving the framework made to ensure that the image detection is accurate and can provide great performance. A quick list of improvements include:
- Allow the ability to support more fonts and letters
- Remove the need for a "grid" on the camera
- Ability to work in dimmer lighting conditions