image_to_text

usage: extract_txt.py [-h] [-v] [-i INFOLDER] [-o OUTFOLDER] [-s]

options:
  -h, --help            show this help message and exit
  -v, --verbose         verbose output
  -i INFOLDER, --infolder INFOLDER
                        input folder containing PDFs to scan
  -o OUTFOLDER, --outfolder OUTFOLDER
                        folder at same level as input folder for output text files. It will be
                        created if necessary. Existing output files with same names will be silently
                        overwritten.
  -s, --skip            skip files already converted and in output directory. Default is to
                        reprocess them and silently overwrite.

This is a thin layer on top of tesseract that processes a folder of images of scanned text and creates another folder of OCR'd text files, using the https://github.com/tesseract-ocr/tesseract OCR engine.

It requires:

• tesseract
• pytesseract
• pdf2image
• PIL

and probably some other things I have forgotten.

Example

You can test functionality of the script with:

./extract_txt.py -v -i test -o output

This will create output with text versions of the four input files located in test.