{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "slideshow": {
     "slide_type": "slide"
    }
   },
   "source": [
    "# Scanned Text Recognition\n",
    "\n",
    "- functionality (mostly done)\n",
    "  - resolution detection, deep language modeling\n",
    "  - font identification, reading order detection\n",
    "  - upsampling, downsampling, better noise removal\n",
    "  - character and word bounding boxes\n",
    "- training / data\n",
    "  - larger, more diverse training sets\n",
    "  - large scale self-supervised training (research)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Scanned Text Recognition (more)\n",
    "\n",
    "- other work\n",
    "  - purely convolutional OCR (better suited to current accelerators)\n",
    "  - replace line normalization, layout extraction with deep models\n",
    "  - replace data augmentation with deep models (like GAN)\n",
    "  - better semantic segmentation (text, image, table, graph, figure, ...)\n",
    "  - non-CTC models and/or automatic decoding\n",
    "  - benchmarking of attention-based models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Camera Captured Recognition\n",
    "\n",
    "- functionality\n",
    "  - page boundary detection models\n",
    "  - DL dewarping\n",
    "  - DL depth estimation (from RGB, from stereo)\n",
    "- training / data\n",
    "  - large collection of photographically captured images\n",
    "  - automatic generation of photographically distorted images (ray tracing)\n",
    "  - automatic DL-based data augmentation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Scene Text Recognition\n",
    "\n",
    "- functionality\n",
    "  - DL text detection / extraction (reimplement standard convolutional models)\n",
    "- training / data\n",
    "  - good datasets exist; train on them\n",
    "  - benchmark CTC vs attention-based models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "celltoolbar": "Slideshow",
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}