Optical Character Recognition: Converting wine catalouge images to text

OCR using Pytesseract

Dependencies:

1.) Pytesseract
2.) Matplotlib
3.) Numpy
4.) openCV

Key features of the ocr-python.py script:

1. Executable Python script that is highly resuable - just add a for-loop to itterate the execution of tesseract on up to as many images as want.

2. Multi-layered filtering- just execute the ocr-python.py script to run multiple filters through images to imporve accuracy of tesseract OCR recognition

Basic Function Filter:

Images Adjusting Filter:

3. Uses matplotlib to allow you to visualize how your filtering functions are actually manipulating images- display is based on a timer that you can control

4. Display the hOCR data by uncommenting this command for each filter

The hOCR data output will look like this:

Next Steps:

1. Enhance image quality
2. Automate rotattion of image
3. Deskewing / Border Removal
4. Cancel noise
5. Logical operation to output the text with the highest confidence
6. Incorporate data training and testing

Next Step Flow Chart: