kanchan88 / Citizenship-Data-Extraction-using-Tesseract

Citizenship OCR using Tesseract

Citizenship-Data-Extraction-using-Tesseract

Just created a SIMPLE OCR which helps to Digitize your Citizenship [ back side of Nepali citizenship as it is in English ]. I used Tesseract-OCR to build this.

What I basically did?

1. Preprocess the data [ converting image to grayscale, blurring image and then adding threshold ]

• We need to grayscale as most OCR use black and white image
• I added thershold of 50 to convert any pixel below 50 to 0 and other to 255. It helps to capture word cleary

2. Sending the image to Tesseract OCR

• Using filters to separate the characters from the background
• Apply contour detection to recognize the filtered characters
• Use mage classification to identify the characters

3. Information Extraction

• OCR gives the raw text. We have to extract useful information from there.
• I used regex and other slicing for extracting information that I need.