nvidia-AI.summerschool17-no.../teaching-lw/033-document-normalization.ipynb

473 lines
2.1 MiB
Text
Raw Permalink Normal View History

2017-09-22 17:26:46 +00:00
{
"metadata": {
"name": "033-document-normalization"
},
"nbformat": 3,
"nbformat_minor": 0,
"worksheets": [
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Title: Document Normalization\n",
"Author: Thomas M. Breuel\n",
"Institution: UniKL"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"\n",
"from pylab import *\n",
"from scipy.ndimage import filters\n",
"gray()"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 73
},
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Example: Grayscale Image Normalization"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(document normalization)\n",
"\n",
"A common problem in document analysis is to turn a stained, poor quality grayscale image back into a binary image.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"doc = mean(imread(\"Figures/historical.jpg\"),axis=2); doc /= amax(doc)\n",
"imshow(doc)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 74,
"text": [
"<matplotlib.image.AxesImage at 0x205d1190>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAFsCAYAAAD7Unz5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZYhkuKkWc0bPojkkWUZ5H/lf
}
],
"prompt_number": 74
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(ink = high value convention)\n",
"\n",
"We generally work on the negative image (ink=high values); this simplifies some of the formulas and boundary conditions.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"doc = 1-doc\n",
"imshow(doc)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 75,
"text": [
"<matplotlib.image.AxesImage at 0x205da950>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAFsCAYAAAD7Unz5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nz1ajgaSlLrPZP8p/I2Cg08uI
}
],
"prompt_number": 75
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"\n",
"\n",
"from scipy.ndimage.filters import minimum_filter,maximum_filter"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 76
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(background estimate)\n",
"\n",
"A good estimate of the paper background is given by the local minimum of the image value.\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"# paper estimate\n",
"paper = minimum_filter(doc,20)\n",
"imshow(1-paper,vmin=0,vmax=1)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 77,
"text": [
"<matplotlib.image.AxesImage at 0x20c06110>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAFsCAYAAAD7Unz5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nL6tSFYfxVGYCRcigev1+fxGH
}
],
"prompt_number": 77
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(shrinkage)\n",
"\n",
"Note, however, that the minimum operation tends to shrink the region, giving fringes. You can see that in this color image where you have green rings invading the background stains.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imshow(transpose(array([paper,doc,doc]),[1,2,0]))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 78,
"text": [
"<matplotlib.image.AxesImage at 0x20c19790>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAFsCAYAAAD7Unz5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nb268cPcelSaA7dvKeUV9nVo3
}
],
"prompt_number": 78
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(opening)\n",
"\n",
"A solution is to compute a maximum after the minimum. This is a special case of _morphological opening_."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"paper = maximum_filter(minimum_filter(doc,20),20)\n",
"imshow(transpose(array([paper,doc,doc]),[1,2,0]))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 79,
"text": [
"<matplotlib.image.AxesImage at 0x20973890>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAFsCAYAAAD7Unz5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n16kO8OPZqKTAltaTSrSnvtQ1
}
],
"prompt_number": 79
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(background subtraction)\n",
"\n",
"We can subtract this to normalize the image partially by paper background value. \n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"deltas = doc - paper\n",
"figsize(12,6)\n",
"subplot(121); imshow(1-doc); title(\"before\")\n",
"subplot(122); imshow(1-deltas); title(\"after\")\n"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 80,
"text": [
"<matplotlib.text.Text at 0x2084ae90>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAADcCAYAAACRZnwYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMSZHmvovISEBAOByuaDT6cLI
}
],
"prompt_number": 80
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(foreground normalization)\n",
"\n",
"Note how the ink is fading in some places; can we normalize that as well? We can perform a local ink-level estimate as well.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"ink = maximum_filter(doc,20)\n",
"imshow(1-ink,vmin=0,vmax=1)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 81,
"text": [
"<matplotlib.image.AxesImage at 0x20334c50>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAFsCAYAAAD7Unz5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZM5OsVhELpcLPBFwythvh91K
}
],
"prompt_number": 81
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(local intensity range scale)\n",
"\n",
"Given the ink and paper values, we can rescale the gray values.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"ranges = ink-paper\n",
"imshow(ranges); print amin(ranges),amax(ranges)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"0.030585106383 0.981382978723\n"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAFsCAYAAAD7Unz5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzsfVmMHNd19tf7MtOzkjMjDpeRRJESJcqWRUlWYiWyLUvO\nQxQ5CCw7AWIkRpCHIC95cPych0R+DbI8OYAQIIYNBLGcBPGaSIYsWwtkbaashRIXcRZyOPv0vvwP\n/E/x9Ol7q25t3dUz9QGD6a6uunXXc797zrnnJjqdTgcxYsSIESNGjBgxQkFy0BmIESNGjBgxYsTY\ny4jJVowYMWLEiBEjRoiIyVaMGDFixIgRI0aIiMlWjBgxYsSIESNGiIjJVowYMWLEiBEjRoiIyVaM\nGDFixIgRI0aICIVsff/738ftt9+O2267Dd/4xjfCeEWMGDFixIgRI8ZQIBF0nK1Wq4WTJ0/ixz/+\nMebn53HffffhW9/6Fu64444gXxMjRowYMWLEiDEUCFyz9dJLL+H48eNYWFhAJpPBl770JTzzzDNB\nvyZGjBgxYsSIEWMokA46wcuXL+PIkSPW98OHD+PFF1/suieRSAT92hgxYsSIESNGjIFCZywMnGyZ\nEqlMJmPdWygUcO+99+LJJ5/E7/zO7xi/J5VKIZFIWH8EVWEbjQbW1tawtraGqakpHDp0yOg9bq2s\n6+vrePHFF/Hd734X5XIZ2WwWjz76KD75yU/iwIEDrtLiSCaTXWVMJBJIJu0VkyZ573Q62Nra8pSn\nf/zHf8Rf/MVf9Fx/5ZVX8P3vfx9vv/12Vx54fhOJBCYmJnD69Gn89m//tnV9dHQUExMTyOVyjvnm\n/03QarXwz//8z/jJT36Czc1NAMD09DQeeeQRfPWrX8Xs7Kz2WepvTu9T/d7pdLC6uopnn30W//qv\n/4r33nsP7XbbNh1q6/vvvx9PPvkkzpw5AwBIp9OYnZ11bH+Zj2QyiVQq1XPfU089ha9//eu2eTHB\nhQsXcO3aNVfPtNttbGxs4G//9m/x0Ucf4WMf+xjuv/9+zM3NIZlMYnR0FIcPH1bKlNHRUdx6663G\n71KlQXKj2Wz2/FatVlEul12Vh5DJZJDL5ZBOuxeviUQC7Xbbtp/VajX87//+L/793/8db731FgBg\nY2MDExMT2mey2SyOHDmCRx55BMeOHcPU1BQymQwKhQJGR0cd89XpdBz7rMS1a9fwH//xH/jRj36E\n1dVVTE9P4/d///fxB3/wBz19kfqnLLebd3700Ud45pln8O1vfxutVstVXoHeMcNljMxHGAqDZrOJ\ndDrtOu10Om3NhQT63G63u+REMpm0/oDrZaO66nQ61l9YkGknEome99LndrttjQW6j4N+k89Suvw/\n/0x1RfVG33Wy264+ms1ml/yo1WraewMnW/Pz87h06ZL1/dKlSzh8+HDPfdls1ip8NpsNOhtKZLNZ\no4nc7zuOHTuGJ554wrp2+PBhlEol5WRnCjkAOcH0OzgymYyn55LJpPLZI0eO4LOf/Sw+8YlP9NzP\nkU6nMT4+jrGxsa5rTiTSD7LZLA4fPozp6WkA14l+v/pfMpnE2NgYjh8/btxmpVIp5FwFAx2Zc3om\nn89jfn4ezWYThUIhpNzpoZrY/I4nuTAKC8lk0iJ09FnXBl6In18Ui0WMjo4im81ak1oqlUI+n1eS\nrXQ63UNq3Ez+IyMjmJ6exqFDh7C8vIxms+mqLWWbDdOxwVLhYHePJCBENGgBF2b/1S0kOGniBCuZ\nTCrJlimZpmdV14Jo33Q63TW2+kq2zpw5g/feew/nz5/HoUOH8O1vfxvf+ta3Akufa7RUmi1dBSYS\nCYyNjaFUKtl2TL8NQJM5EUyahDKZTKAkgspgtwq2ux6GIKH2uOmmmywyQ3l1GsC8/UwJV6fTQbPZ\ntO3g8h133nkn5ufnuwbrwYMHjZ73g1QqhYMHD+K3fuu3XD2Xy+V8aUT7hXw+74kYZjIZPPTQQzh5\n8iSKxSJmZmZQKpWQTCaRy+WsFajEIMhDlFAsFnHrrbdiZGQEAPDmm2/i9OnTPRoOAmkKi8ViX904\nkskkRkZGcPDgQUxNTaFYLHZpVvh9JoTBDqVSCXfeeSc2Nzfx+uuvY3FxETs7O560XMANOVmv1z3n\nadBQWUNUC3fgxlyl60NBgGuBuMznf61Wq0uzRXnjRJzPf/y6k1Z4kAhcYqXTafzDP/wDHnvsMbRa\nLXz1q18NfCciTcbUcei/nWkpnU53abRU6umgmG6xWLS+h9l5dSYQgq487XbbswDieOCBB7q+08pa\nag6JGNNnHZxWGyo1f71e71qZOeE3f/M3ldedyDD1M6/1ls/ncfz4cRw/ftz1s9ls1tIg+tGOqvCp\nT30qkHRKpRLy+bzx/dSWlUoF999/P2q1GprNZte4JG0HkS4Or9pYXT44dCbHKGF+fh6f/vSnre93\n3XUXTp8+7fhcIpFAPp/HyMgI0uk0stlsj2ZXNc6o37s1Jc7MzOATn/gEEokERkdHccsttyCfz/fU\nJ030fjRbk5OTuPvuuzExMYHjx4/jrbfewtWrV9FoNFzlmb+7Xq/j8uXLuHLlCqrVaqgTuXSz8AOp\nsbLTWtHvcq7iZQ26//N8qMgWNyHS+3l+iIy1Wi2LnAE35jZuYowKAg/9YPTSRAIjIyNdPlunT582\n8tlKJBLI5XI9ZIvDpEg6oaGz2/tBWGaxIMhWp9NBpVIJJD+5XE5pIiCYqLndki0+QIMmIjJviUTC\ncZWrU5EDwfUDEoh22kKVz1aY5lkvgq3dbmNzc9OaECuVSpeWMplMYmZmRpnvfD6Pubk543e59dmq\n1+uoVqu2z/NycGQyGWQyGU/90dRnq91uK81Cpu+gvKneQ+XhZp1Go4FWq2URF5P2rtVqqFQqaDab\nSCQSlhvH9PS0kmyRyYjDTb8iTXez2USj0cDi4iJ2d3ddE0RCs9nE9vY2nn/+efz0pz/FxYsXA1mk\nOsHtWCWiyuuUFiqdTqfLT0vny8R/V/luBW2V4Xmh9qF+L8kSJ44c1C+5ubjVaqHZbFokjMYJLzOV\nNZVKdfm7efHZktjc3NTevyd08aZajRh7E2ERCdVAM+lnYa5f5PtTqVRoJMrrJOUWcuVNsCMrYY93\nlcZHhXa7bTnSB2EKcwud36TTM9xBWgciWETO2u225VdXrVYd+0c2m+0xLYdpokokEhbRLRQKXb6g\nbkATfq1WQ7lcxs7ODs6dO4erV6+iXC6HrjFRKRDsoKpTajNJtoAbRExF1rnPlHxHUODjnb+TyDb1\nNU76VGSICJXKB01FKgeNPUG2YsQIA1IQkR+RHdrtNur1eqTU11FHsVhEq9VS7orTCcswtXQxYugg\niUkUYOeHRb9LVxuTNMO0yEgtG3eE55CO8vTXbDa75Cz9p8Vns9ns8vOSxG0QiMnWkCJKjD3GDZC5\nRIdOp+PZh4Se34sYVH8O8r39LMNeHP9uyEA/oNJqk7aIJvwoQGUeJKJERIU2lJAGiZvOpP8UNyWa\nQmemVxE2Tv5UmnqZBqXDNan1et0qBzc/kztJOp3uMkdS2YPyV/aCPUO29pspcb+Vd5gQZrvEbR4s\n/E7unFzzCS/MdiKz4aAmjbAQJaIF6MdaVEghkT9OljgB4T7N5PxO5IrMiNz3D9CbEZ3yYVdXqms6\nsqWLJynvzeVySKVSqNfrFrECbpgWpeM8mYa5T1e/ZenAyBbfuRTUNu6YgMTY64iCkI/RjTA3ZqhA\n2oJBbnPvt2/aIMDrtlQq4dChQ10bi7impdFoYGdnB7u7u740127A/RyJkHCnb94veRw8rnFSkRvA\nrG85bdKxe87EL407x8v7eT51voOcQHJuMKhdigMjW9z3pd/Cyg28CpSgt/EC3RGNgzZ/BNX5dCsZ\n3RZ7vrpyk88oOkCaIKi+IL9HZbXtFul0GoVCoSdWjtOuT6+TfRTqSY7nQeaDT1om9xNUAUjdvlOX\nPp8PVLsT3b7XzsnaKa/A9fAV
}
],
"prompt_number": 82
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(result of normalization)\n",
"\n",
"We can now normalize this range so that ink = 1 and background = 0 locally across the whole image.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imshow(1-deltas/ranges)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 83,
"text": [
"<matplotlib.image.AxesImage at 0x20728910>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAFsCAYAAAD7Unz5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBIbiB4MB2u02Go2GJcM1HA4R
}
],
"prompt_number": 83
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(thresholding)\n",
"\n",
"We can now _threshold_ the image to get back the printed ink.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imshow(deltas/ranges<0.5)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 84,
"text": [
"<matplotlib.image.AxesImage at 0x20d3ead0>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAFsCAYAAAD7Unz5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxQssXbqUDnEOGTKE9ryUZd26
}
],
"prompt_number": 84
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(detecting background-only regions)\n",
"\n",
"This obviously doesn't work well in areas where there is no ink at all. We can fix this by requiring a minimum range difference; if that isn't met, we assume that the local region contains just background (paper) pixels.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imshow(where(ranges<0.3,1.0,deltas/ranges<0.5))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 85,
"text": [
"<matplotlib.image.AxesImage at 0x2049f190>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlsAAAFsCAYAAAD7Unz5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nOGVUvemIRCISGRkptU58Pp+M
}
],
"prompt_number": 85
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(normalized gray level image)\n",
"\n",
"We can also derive a reasonable normalized gray level image with a little bit of arithmetic.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"normalized = deltas/(maximum(0.7,ranges))\n",
"subplot(121); imshow(1-doc); title(\"before\")\n",
"subplot(122); imshow(1-normalized**1.1); title(\"after\")"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 86,
"text": [
"<matplotlib.text.Text at 0x201f5110>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAADcCAYAAACRZnwYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\neKLK67dNBH9Vzp47d26//iB9
}
],
"prompt_number": 86
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 86
}
],
"metadata": {}
}
]
}