nvidia-AI.summerschool17-no.../teaching-lw/035-binary-morphology.ipynb

674 lines
399 KiB
Text
Raw Permalink Normal View History

2017-09-22 17:26:46 +00:00
{
"metadata": {
"name": "035-binary-morphology"
},
"nbformat": 3,
"nbformat_minor": 0,
"worksheets": [
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Title: Binary Morphology\n",
"Author: Thomas M. Breuel\n",
"Institution: UniKL"
]
},
{
"cell_type": "code",
"collapsed": true,
"input": [
"\n",
"from scipy.ndimage import filters\n",
"def imrow(*args,**kw):\n",
" size = kw.get(\"s\",8)\n",
" if \"s\" in kw: del kw[\"s\"]\n",
" n = len(args)\n",
" gray()\n",
" subplots(1,n,figsize=(n*size,size))\n",
" for i,im in enumerate(args):\n",
" subplot(1,n,i+1); imshow(im,**kw)\n",
"\n",
"test = mean(imread(\"testimage.jpg\")/255.0,axis=2)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 14
},
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Binary Morphology"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(mathematical morphology)\n",
"\n",
"Another useful set of non-linear filters comes from *mathematical morphology*.\n",
"These filters are easiest to describe on binary images, so that's what we are starting with.\n",
"\n",
"Generally, given an image containing shapes (identified by pixels with value 1), \n",
"in binary mathematical morphology, we try to fit a second shape (the mask)\n",
"into that shape somehow. This is illustrated by the following images denoting the four\n",
"basic operations of mathematical morphology: erosion, dilation, opening, and closing.\n",
"\n",
"![Erosion](files/Figures/erosion.png)\n",
"![Dilation](files/Figures/dilation.png)\n",
"![Opening](files/Figures/opening.png)\n",
"![Closing](files/Figures/closing.png)\n",
"\n",
"(images from Wikipedia)"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"# circular structuring elements\n",
"from scipy.ndimage import morphology\n",
"xs,ys = meshgrid(range(20),range(20))\n",
"m3 = ((xs-10)**2+(ys-10)**2)<=9\n",
"m8 = ((xs-10)**2+(ys-10)**2)<=64\n",
"imrow(m3,m8,interpolation='nearest',s=4)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAdsAAADlCAYAAAAbfV1dAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n585yfHy8/P333+te5qr8/+/l
}
],
"prompt_number": 15
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"# morphologial dilation\n",
"imrow(test,morphology.binary_dilation(test,m8))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAA6EAAAFOCAYAAAB31htUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhQdt2rlzJ3M6nWpfHT16lEVH
}
],
"prompt_number": 9
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"# morphological erosion\n",
"imrow(test,morphology.binary_erosion(test,m3))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAA6EAAAFOCAYAAAB31htUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n227DnnvuGXMti0t7gHR/33DD
}
],
"prompt_number": 10
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"# morphological opening\n",
"imrow(test,morphology.binary_opening(test,m3))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAA6EAAAFOCAYAAAB31htUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n8CYnZQA+NJeJD83r+u///m+K
}
],
"prompt_number": 11
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"# morphological closing\n",
"imrow(test,morphology.binary_closing(test,m8))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAA6EAAAFOCAYAAAB31htUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkIxz9erVWLx4ccTbP/XUU0nU
}
],
"prompt_number": 12
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(mathematical definitions)\n",
"\n",
"Mathematically, these operations can be defined in terms of sets.\n",
"Let the pixels of the image be given by the set $A$\n",
"and those of the mask by the set $B$.\n",
"\n",
"Denote by $B_z$ the set of translated points $B_z = \\\\{b+z | b\\in B\\\\}$.\n",
"\n",
"- Erosion: $A \\ominus B = \\\\{z \\in E | B_z \\subseteq A \\\\}$\n",
"- Dilation: $A \\oplus B = \\\\{z \\in E | B_z \\cap A \\neq \\oslash \\\\}$\n",
"- Opening: $A \\circ B = (A \\ominus B) \\oplus B$\n",
"- Closing: $A \\bullet B = (A \\oplus B) \\ominus B$\n",
"\n",
"The set operations are as usual. \n",
"We also need reflection of a set of points through the origin,\n",
"$\\hat{B} = \\\\{-x | x\\in B\\\\}$"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(Python equivalents of other oprations)\n",
"\n",
"\n",
"Incidentally, remember how we can implement these binary operations on images:\n",
"\n",
"- $I^c$ = `1-image`\n",
"- $\\hat{I}$ = `I[::-1,::-1]`\n",
"- $B_z$ = `roll(roll(B,z[0],0),z[1],1)`\n",
"- $A\\cap B$ = `minimum(A,B)`\n",
"- $A\\cup B$ = `maximum(A,B)`"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(minimum / maximum filter equivalents)\n",
"\n",
"Erosion and dilation can also be thought of as _local minima_ and _local maxima_ filters over binary images, defined over a region determined by the mask $B$.\n",
"\n",
"The minimum and maximum filters from the library effectively compute morphological operations over rectangles."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(mathematical properties)\n",
"\n",
"There are also some mathematical properties these operations satisfy,\n",
"somewhat similar to what we found for linear filters (not exhaustive):\n",
"\n",
"_translation invariance_ or _shift invariance_\n",
"\n",
"duality:\n",
"\n",
"$$A \\oplus B = (A^c \\ominus \\hat{B})^c$$\n",
"$$A \\bullet B = (A^c \\circ \\hat{B})^c$$"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(properties of dilation)\n",
"\n",
"commutativity: \n",
"\n",
"$$A \\oplus B = B \\oplus A$$\n",
"\n",
"associativity: \n",
"\n",
"$$(A\\oplus B)\\oplus C = A\\oplus (B\\oplus C)$$\n",
"\n",
"distributivity over set union: \n",
"\n",
"$$A\\oplus (B\\cup C) = A\\oplus B \\cup A \\oplus C$$"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(properties of erosion)\n",
"\n",
"NB: these are not the same as dilation\n",
"\n",
"alternative to associativity:\n",
"\n",
"$(A\\ominus B)\\ominus C = A\\ominus (B\\oplus C)$\n",
"\n",
"distributivity:\n",
"\n",
"$$A \\ominus (B \\cup C) = A\\ominus B \\cup A \\ominus C$$"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(idempotency and monotonicity)\n",
"\n",
"\n",
"opening and closing are _idempotent_: \n",
"\n",
"$$(A \\bullet B) \\bullet B = A \\bullet B$$\n",
"\n",
"monotonicity properties:\n",
"\n",
"$$(A \\ominus B) \\subseteq (A \\circ B) \\subseteq A \\subseteq (A \\bullet B) \\subseteq (A \\oplus B)$$\n",
"\n",
"if origin is in $B$ and $B \\subseteq C$ :\n",
"\n",
"$$A \\oplus B\\subseteq A\\oplus C$$ "
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"d = zeros((12,12))\n",
"for i in range(12): d[i,i] = 1\n",
"imrow(d,morphology.binary_erosion(test,d),morphology.binary_opening(test,d),interpolation='none')"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABVgAAAGkCAYAAADE5U6gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQEECVgAAAACAggSsAAAAAAAF
}
],
"prompt_number": 13
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"As you can see, this kind of \"detects patterns\" in the image."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Hit or Miss Transform\n",
"=====================\n",
"\n",
"There is a useful combination of these operations called the *hit or miss transform*.\n",
"\n",
"$A \\odot (B,C) = (A \\ominus B) \\cap (A^c \\ominus C)$"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"b = zeros((12,12))\n",
"for i in range(12): b[i,i] = 1 \n",
"c = zeros((12,12))\n",
"c[11,0] = 1\n",
"c[0,11] = 1\n",
"import numpy\n",
"imrow(bitwise_and(morphology.binary_erosion(test,b),morphology.binary_erosion(1-test,c)))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAeIAAAFsCAYAAAAQU7oLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nu+/WqFGjlJaWpj//+c+Svrsy
}
],
"prompt_number": 14
},
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Hit-or-Miss Example"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's say we want to find the letter \"I\" in this image.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"caps0 = 1*(mean(imread(\"caps.png\"),2)<0.5)\n",
"gray(); figsize(12,6); imshow(caps0)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 44,
"text": [
"<matplotlib.image.AxesImage at 0x9b2eb50>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAArsAAABVCAYAAABJsixZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nm1+yZAkOHTrEWzZx4kTR2zdF
}
],
"prompt_number": 44
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Usually, images are noisy and degraded. We can simulate that with this simple document image degradation model. Noise means that we can't do pixel accurate matching.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"caps = 1-mean(imread(\"caps.png\"),2)\n",
"caps -= amin(caps); caps /= amax(caps)\n",
"caps = filters.gaussian_filter(caps+0.0,1.0)\n",
"caps += 0.4*(rand(*caps.shape)-0.5)\n",
"caps = 1*(caps>0.5)\n",
"imshow(caps)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 47,
"text": [
"<matplotlib.image.AxesImage at 0xa941d90>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAArsAAABVCAYAAABJsixZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkQaL8tOwYUPROsvLywX7XL9+
}
],
"prompt_number": 47
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"subplot(121); imshow(caps0[:,:100])\n",
"subplot(122); imshow(caps[:,:100])"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 48,
"text": [
"<matplotlib.image.AxesImage at 0xabf20d0>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAArsAAAEQCAYAAACwULwyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n3fbMXltbkx4fHh5GcnIyZmdn
}
],
"prompt_number": 48
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Here is the template for the letter.\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"template = 1*(mean(imread(\"I.png\"),2)<0.5)\n",
"gray(); imshow(template)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 49,
"text": [
"<matplotlib.image.AxesImage at 0xb069ed0>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAKAAAAFpCAYAAAAbXyPlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nigGSKAZIohggiWKAJIoBkigG
}
],
"prompt_number": 49
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We require most of the pixels to be present. Since the edge of the character may be degraded, we are using a slightly smaller template to match the inside pixels. We obtain that template using mathematical morphology.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"inner = morphology.binary_erosion(template,iterations=1)\n",
"imshow(inner)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 50,
"text": [
"<matplotlib.image.AxesImage at 0xaf50a90>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAKAAAAFpCAYAAAAbXyPlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJFEMkEQxQBLFAEkUAyRRDJBE
}
],
"prompt_number": 50
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"When we match with this template, we get a large number of false matches and one true match.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imatch = morphology.binary_erosion(caps,structure=inner)\n",
"imshow(transpose(array([imatch,imatch,caps]),[1,2,0]))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 57,
"text": [
"<matplotlib.image.AxesImage at 0xc3b0990>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAArsAAABVCAYAAABJsixZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xl4FFXe6PFvIAFkkZ0gCRBkMSEiBFlGR2VR4Loksl02\nH8zL5syr8ypeB1FnXq+8KoLLzChefXGug7wyI+BVFh1gEEH0dQEx7uigkgxhi8OmEJCQcO4fJ5VU\nn67urq6qDiH8Ps9zHujq6nNq68rpU+f8TpJSSiGEEEIIIUQdVO9Mb4AQQgghhBCJIpVdIYQQQghR\nZ0llVwghhBBC1FlS2RVCCCGEEHWWVHaFEEIIIUSdJZVdIYQQQghRZ/mq7K5bt47MzEy6d+/O/Pnz\ng9omIYQQQgghApHkNc5uRUUFF110ERs2bCAtLY3+/fvz0ksvkZWVFfQ2CiGEEEII4Ynnlt2tW7fS\nrVs3MjIySElJYcKECaxatSrIbRNCCCGEEMKXZK8f3LNnDx07dqx6nZ6ezpYtW0LWSUpK8r5lQggh\nhBBCuBSps4Lnll2pyAohhBBCiNrOc2U3LS2N4uLiqtfFxcWkp6cHslFCCCGEEEIEwXNlt1+/fnzz\nzTcUFRVRVlbGsmXLyMvLC3LbhBBCCCGE8CVmZXfq1KmkpqbSq1evqmWHDh3i2muv5eTJk2RnZ3PR\nRRcxfvx4icQghBBCCCFqlZiV3SlTprBu3bqQZfPmzWPYsGHs3r2b+++/n/Hjx3PvvfcmbCOFEEII\nIYTwwlWc3aKiInJzc/n8888ByMzMZPPmzaSmprJ//34GDx7M119/HZ65DGITQgghhBA1INBoDCUl\nJaSmpgKQmppKSUmJ9y0TQgghhBAiQXxNFwy69VZacIUQQgghRG3kqbJrdV8A2LdvH+3atQt0o4QQ\nQgghhAiCp8puXl4eixcvBmDx4sWMHDky0I0SQgghhBAiECqGvLw81aBBAwWo5ORkNWnSJHXw4EF1\n1VVXqfPOO081btxYDRkyRB0+fDjss4AkSZIkSZIkSZIkSQlPkcRs2V24cCFbtmxBKcXhw4fZtm0b\nJSUlDBw4kAceeIDS0lJGjBjBvHnzYmUlhBBCCCFEjYpZ2W3fvj19+vQBoGnTpmRlZbFnzx5Wr15N\nfn4+APn5+axcuTKxWyqEEEIIIUScXMXZtRQVFTFo0CC++OILOnXqxOHDhwFQStGqVauq11WZS5QG\nIYQQQghRA3zH2T127BhjxozhySefpFmzZiHvSfgxIYQQQghRG7mq7J46dYoxY8YwefLkqsgLEn5M\nCCGEEELUdsmxVlBKMW3aNHr27MnMmTOrllvhx2bPnh1Q+LGmwK0Oy/8C7PaZt+XuKO/tRNf9lwdU\nViT/AmwEJtiWPeozz/OBX0Z4z2/e0UQ7nk6WAf/wUM4dQEOX6y4C/umhjGic9vMroAXwos+8LwJu\ntL0+BjwT4zNNgNtsr58Djngo275fnwJ/85CHmc/nwFoPeVwD9LW93g68HmHdAcBgY9lKYEecZV4K\nXG17HfR35Tzg34xl8ZThdN153cbhQB9j2WIg0uyXt6GvM4Bi4CWP5Tox9+s/0fevmrrXW75H7393\n4KmAyrVcDFxnLFsPfAFcDrwdUDm3ov92RvIhkAk8G1B5pmjH2eu1Gu/fFdP7wDtxrJ8PpFb+/zDw\nxxjr27dvK/BWHGWZn9+EPkdu9EXfJ0F/H4tdfu469PVoiXVeZgINKv9fCLwcZd0pQFuX+dolA//L\nWBbr85OBC2yv3wQ+cl9ktLBjJ06cUFlZWQpQDRs2VO3atVN9+vRRy5cvV4MGDaoKPTZ48GCfocee\nVFCkQDmkKwMIRzFEwbcR8rdSqYLjlet9G3A4jNts+f6kYI9RtvVeKw95PxPl2NnztlJSQPv0VYzj\n6ZSu9lDOHQoq4ihjV4DnzTpmTuUcrTyXfq+XsUa+e1x8ptD4TFePZdvzWOgxj4eNfP7kMZ+njXyW\nRFivvoIDDudjn4cy/2nkEdR1Y6U2Dtv5rYLLPZwfpaCPx+3oouAHh/xyonxmv229EwqGB3A8In2f\niir/3V35fm4AZbm5V5xU8KPS95dvFTQOoFzrXl/iUN73Cr5TcKhyncd9ljVDwakY+3i4cp0FAeyb\nPV0c4Vya1/ooD3nf5/L8RUoPxlneR8Y2x3NtPRZnWcuNzx9QUM/lZ+33Kzd/J6z0glHmQxHWO79y\n/+1/b9dFyfdbpb9D1rrr49imRg7n7VsF10Qp6ydj/Tsc1/UUeqxRo0Zs27YNpRTHjh2jS5cuLFiw\ngA8//JDrrruO48ePc//99zNw4EBatGgRLasYUoHOEd7r5yNfSxOga4x1GqNbYrpWJgX8b5/lZlXm\n87Qt34ZAB2M9672DHsqIduzseVvpdOU23eShLLsLfX7erZbEN/dJxwDKvADdSmgdMydN0efSWqco\ngHLdygggj0UB5AHVv+oT7TxgGFAOtDbeewP4nx7ybON3ozzoCvw8xjqtgW0Oyz/2WGYK+umPaaDL\nzzeiupXXi1T0uYv0fbLuX2mV7/8M2OujPLcaAM3Q95euQCnQ00d+uVTf65269bVF3zdbVq5zF5Dj\no7wWxH44a63zK/R9f4CP8kDvw6foJzix/qZ2BV71UMbDHj5jqQC+8/H5ruhj5davCf97Hs1edGuw\npTV6m2MpI/R+Fc/341+M179BP1E0Wd8D+9/bEcA9EfJdSXULMLhvaQb4CbjPWNYVfS83XUV13clS\nBDwZR3kuahGNGzcGoKysjIqKClq2bBlw2LHuRL/5/85H3n6NQ/+R9aIDkR/D1gZL0DfbmvIR+gZZ\nm41CP7JZib4uRe3xCvoxsOle9CP6/67ZzfEl1uO6/uguFonm9tH2V8AWH+WsxPncRXIfukL4MlDf\nR7lezIy9iqNfAKs9fO5R4AGPZXqxFt2NwqufA5cEtC2J8BPwwpneiChmorvl+fXrAPJwy+mH7mLC\n6w/T4sx3BbEbKUbi3L3OqctrdDEru6dPn6ZPnz6kpqYyZMgQsrOzKSkpITVV93FJTU2lpCRSvy83\ndgDpMdZp7iN/Jwp9opJsaRnhvypeQ3954tUM3cJotn6eBHoZ5SahW04Ooltd/XrcIf9Flfmb+zKI\n8FYyr250KNee+qH7yPn1txjl+IkKkgmMJbT14yj62B10KOcN23siWA3RLVIpwF+Bax3WSQLO1sls\nPo3yXrMo7wVNAXkx1snCfSuwXSNgD7ql1nIUaEX4d8n8HjVCfxeDivLTJEKZJx22Od4KdmvCK5An\n0f0krbK6AVMdPnsNunL/P+Is00k5+lxaZdYDjhvr7MT5qYFXp9HHsSGhx3ZV5fZ4uTc63dOdWvGO\noH8E29eL1n+5Nov25GQe+j5oKQNOJWAb7KFjj9n+73SvutN4fdRDeV8D/88oC/Q96VfofR4C/Gh7\n7xT6mop/XEjMym69evX45JNP2L17N2+//TabNm0KeT/4sGOPOCxzatr24zDhLcYTHNabjbcvzyjg\nA2PZ79FdB75wWP8k+hHFv3soy42plfmPNpbnElzLuZfHyGeDm9HHzumR93Dbe0/X5EadAzLRj4Iv\nI3ygD8T7CKv2aUfkFrJED5KtKbcR/oh3OnCLw7rW98jsXnNvArbLXuYkY9lkoEec+RxA3yfsbgK+\ntL3+Dr1vDxM+oOldYF2cZTrZgm6gsSjC/6b0Q7eWeTXWeP1b9HEsM5aPRFeIguoq5NTifqXD9pwN\n9qAH4NrF03VjLfBenGX+3ngdaUC7ZQj6ugb9hMX8zppdRZzqT279a4Tl9wC3E9ol6AO8XlOuO0M2\nb96c66+/no8++ijAsGNO/XluAuZ4zC9ov0f344rXUOP1AvRjh1difG6uh7LisZbw/r0jAsr7+YDy\nMdVU/OZ/ED4KuDPuIxQ8HuzmnPO6oEdVmxW/cvR5qckuOInQntDoE5Y3a3pDgP+bgDyT0BUhu4/R\nXbvmR/ncfxqv/yPIjXLgpU9p
}
],
"prompt_number": 57
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We need a second template that asserts that the outside of the character should not contain pixels. We also construct this using mathematical morphology.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"outer = maximum(0,morphology.binary_dilation(template,iterations=4)-morphology.binary_dilation(template,iterations=2))\n",
"imshow(outer)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 52,
"text": [
"<matplotlib.image.AxesImage at 0xb546f90>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAKAAAAFpCAYAAAAbXyPlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nt1yA6hpPzeepOT61to8E2B4B
}
],
"prompt_number": 52
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can visualize the relationship between the outer and the inner templates by combining them numerically.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imshow(inner-outer)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 53,
"text": [
"<matplotlib.image.AxesImage at 0xb762990>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAKAAAAFpCAYAAAAbXyPlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUlXAEydOIJlMYuvWrfZr9+/f
}
],
"prompt_number": 53
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can now visualize the combination of the `imatch` and `omatch` locations.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"omatch = morphology.binary_erosion(1-caps,structure=outer)\n",
"imshow(transpose(array([omatch,imatch,caps]),[1,2,0]))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 59,
"text": [
"<matplotlib.image.AxesImage at 0xc84d450>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAArsAAABVCAYAAABJsixZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\no4RDoS9Uji09jfur4g8EmWqZ
}
],
"prompt_number": 59
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The actual match location is where both the inside and the outside match. This is the hit-or-miss transform.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imshow(transpose(array([imatch*omatch,zeros(omatch.shape),caps]),[1,2,0]))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 62,
"text": [
"<matplotlib.image.AxesImage at 0xd4e2710>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAArsAAABVCAYAAABJsixZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIbt30UO7VoV5338Yj//dz0o5
}
],
"prompt_number": 62
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": []
}
],
"metadata": {}
}
]
}