Intelligent Character Recognition.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Intelligent Character Recognition\n",
    "(AKA ICR)\n",
    "\n",
    "See:\n",
    "- https://en.wikipedia.org/wiki/Intelligent_character_recognition\n",
    "\n",
    "This is not the same as OCR.\n",
    "\n",
    "> When we talk about OCR, we are really talking about converting data from an image that was created by a machine, whether it be a document created by an office application, and even an old document typewritten\n",
    ">\n",
    "> For ICR, pattern-matching goes to a different level. You cannot really have success deploying a solution requiring ICR without additional information to aid in the recognition process.\n",
    ">\n",
    "> https://www.parascript.com/blog/difference-ocr-icr/\n",
    "\n",
    "# Segmentation challenges\n",
    "\n",
    "A few common segmentation methods are listed below:\n",
    "![Common segmentation methods](img/common_segmentation_methods.png)\n",
    "\n",
    "See: https://www.scanstore.com/ICR_Guide/\n",
    "\n",
    "## Possible solutions\n",
    "\n",
    "1. Use the common pattern to detect text field (AKA Handprint Recognition)\n",
    "\n",
    "2. Use imagemagick to remove the form/table structure:\n",
    "\n",
    "```\n",
    "convert inputFile.jpg -type Grayscale -negate -define morphology:compose=darken -morphology Thinning 'Rectangle:15x1+0+0<' -negate outputFile.jpg\n",
    "```\n",
    "\n",
    "Advantages:\n",
    "- Remove most horizontal and vertical lines\n",
    "- Clearer text (remove unnecessary elements on page)\n",
    "\n",
    "Disadvantages:\n",
    "- It introduces extra spaces (becomes monospaced font)\n",
    "- It can remove some part of handwriting text\n",
    "- Hard to determine the perfect rectangle size (vary: 15)\n",
    "\n",
    "# Let's explore the available options on the market\n",
    "\n",
    "## queXF\n",
    "An open source, web based paper form verification and data entry system\n",
    "\n",
    "> The process of ICR is broken in to the following 7 steps:\n",
    "> - Character isolation\n",
    "> - Noise reduction\n",
    "> - Boundary removal\n",
    "> - Normalising\n",
    "> - Thinning\n",
    "> - Feature extraction\n",
    "> - Training or Recognition\n",
    ">\n",
    "> https://quexf.acspri.org.au/intelligent-character-recognition\n",
    "\n",
    "It can:\n",
    "- Automatically detects locations of boxes on the form (no need to manually overlay boxes)\n",
    "\n",
    "\n",
    "## NIST FORM-BASED HANDPRINT RECOGNITION SYSTEM (RELEASE 2.2)\n",
    "A public domain document processing system was developed by the National Institute of Standards and Technology (NIST) in 1994. The system is a standard reference form-based handprint recognition system for evaluating optical character recognition (OCR), and it is intended to provide a baseline of performance on an open application. \n",
    "\n",
    "NIST developed a new release in 2003 for internet distribution of its standard reference form-based handprint recognition system for evaluating optical character recognition (OCR). Release 2.2 incorporates all bug fixes generated for previous versions.\n",
    "\n",
    "https://www.nist.gov/services-resources/software/public-domain-ocr\n",
    "\n",
    "For source code and training data:\n",
    "- https://github.com/brainopener/nist-ocr\n",
    "\n",
    "According to the documentation:\n",
    "- To process a new type of form, trainreg must be run on a prototypical form and the output coordinate stored.\n",
    "\n",
    "##  ABBYY Cloud OCR SDK\n",
    "\n",
    "https://ocrsdk.com/documentation/quick-start/text-fields/\n",
    "\n",
    "## Others\n",
    "- https://www.a2ia.com/fr/a2ia-fieldreader\n",
    "- http://www.recogniform.com/index.htm\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}