OCR Text Reader From Camera
Features
- Take a picture with your device camera
- Uses TesseractJS to scan for text
- Output the registered text and logging
A simple app that uses the camera widget to take a picture and process out the text using the Tesseract library.
I tried to play with Tesseract one or two years ago and literally fell into a rabbit hole. The image quality is very-very condition sensitive and requires tons of pre-processing and maybe a PhD in Computer Vision.
From what I learned the main goal is to get as contrast and sharp text as possible (black readable letters on white background) so that tesseract's neural network can confidently recognize them. So, in many applications OpenCV library does some heavy lifting even before the text gets into tesseract.