nvidia-AI.summerschool17-no.../teaching-lw/070-template-matching.ipynb

1169 lines
1.4 MiB
Text
Raw Permalink Normal View History

2017-09-22 17:26:46 +00:00
{
"metadata": {
"name": "070-template-matching"
},
"nbformat": 3,
"nbformat_minor": 0,
"worksheets": [
{
"cells": [
{
"cell_type": "code",
"collapsed": true,
"input": [
"\n",
"from pylab import *\n",
"from scipy.ndimage import filters\n",
"import os\n",
"def imrow(*args,**kw):\n",
" gray()\n",
" size = kw.get(\"s\",6)\n",
" inter = kw.get(\"i\",\"nearest\")\n",
" if \"s\" in kw: del kw[\"s\"]\n",
" if \"i\" in kw: del kw[\"i\"]\n",
" n = len(args)\n",
" subplots(1,n,figsize=(n*size,size))\n",
" for i,im in enumerate(args):\n",
" subplot(1,n,i+1); imshow(im,origin='upper',interpolation=inter,**kw)\n",
"def rgb(r,g,b=None):\n",
" if b is None: b = zeros(g.shape)\n",
" return transpose(array([r,g,b]),[1,2,0])\n",
"def N(a): return a/max(amax(a),abs(amin(a)))"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 12
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Template Matching\n",
"=================\n",
"\n",
"Let's start with a seemingly simple problem: finding a letter on a page.\n",
"We assume that we have a template for the letter, and an image in which\n",
"the letter is at exactly the same scale, and in which it is present\n",
"in exactly the same font."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"template_A = mean(imread(\"simple-ocr/template-A.png\"),axis=2)\n",
"textimage = mean(imread(\"simple-ocr/small1.png\"),axis=2)\n",
"print template_A.shape\n",
"imrow(template_A,textimage[58:87,30:59],textimage)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"(29, 29)\n"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABAoAAAFBCAYAAAD+A00kAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n1d/PY39Prme63wOx7y8zk/cw
}
],
"prompt_number": 5
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"A simple idea for finding the letter in the image is the following.\n",
"We slide a window across the image that is the size of the letter.\n",
"Within the window, we compute the squared (or absolute) difference\n",
"between the letter templat and the image.\n",
"The letter is present wherever that distance is small.\n",
"To visualize that better, we can also display the inverse\n",
"distance image and"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"tw,th = template_A.shape\n",
"w,h = textimage.shape\n",
"ds = zeros((w-tw,h-th))\n",
"dw,dh = ds.shape\n",
"da = zeros((dw,dh))\n",
"for i in range(0,dw):\n",
" for j in range(0,dh):\n",
" ds[i,j] = sum((template_A-textimage[i:i+tw,j:j+th])**2)\n",
" da[i,j] = sum(abs(template_A-textimage[i:i+tw,j:j+th]))\n",
"imrow(ds,da)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAB7CAYAAACVf11OAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n1hrXwQ7AO3AbUsw6K5fLkhTk
}
],
"prompt_number": 18
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"It's a little easier to see what's going on in the negative image.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(-ds,-da)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAB7CAYAAACVf11OAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nyT7/Xl9fS1KAd56HbUfmscP0
}
],
"prompt_number": 19
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Because the minimum distances aren't easy to identify, let's display this in a different way."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"max1 = filters.maximum_filter(1.0/(5.0+ds),(10,10))\n",
"max2 = filters.maximum_filter(1.0/(5.0+da),(10,10))\n",
"imrow(max1,max2)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAB7CAYAAACVf11OAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nU4FbGNtawDEermrsisBxjD6v
}
],
"prompt_number": 20
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Statistically, it's more important to look at the distribution of distances."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"_=hist(ds.ravel(),bins=100)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAD9CAYAAACx+XApAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nbm4mmUwSiUQIh8P4/X7y8vLI
}
],
"prompt_number": 21
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"From this, we can read off that a threshold of 60 might be pretty good for finding\n",
"the locations of these letters.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(60-clip(ds,0,60))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAACECAYAAACNmKC5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n1dTUZGp6R9XU1Ki9vV2S1N7e
}
],
"prompt_number": 22
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Another way of automatically characterizing a distribution like this is by\n",
"mean and variance.\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"mean(ds),sqrt(var(ds))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 41,
"text": [
"(103.4008935505834, 14.741544406032297)"
]
}
],
"prompt_number": 41
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Outliers in are pixel values that are \"far from\" the mean,\n",
"where \"far\" is defined in terms of the variance.\n",
"(Like \"IQ\")"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"_=hist(((ds-mean(ds))/sqrt(var(ds))).ravel(),bins=100)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAD9CAYAAAC/fMwDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAHCRJREFUeJzt3X1sW9X9x/G3WYw2RtOB1jjI7rB+jaPUJYXw4GaT2MyK\n0yysURldtICalAKa0j0ENvHQf0aLRGOG0NaWZRMoIC9IpNn+aDJUvFANM2DDZWkZ04wabwsidhwP\nFtIGCJi29/dHyG3Th9RpHpzkfl5SJefYJ/5e1fnk5Nxzz7UZhmEgIiIL3gW5LkBERGaHAl9ExCIU\n+CIiFqHAFxGxCAW+iIhFKPBFRCxiwsA/dOgQZWVl5r/Fixezc+dOBgcHCQQCFBcXU1FRwdDQkNmn\nqakJj8dDSUkJXV1dZnt3dzelpaV4PB4aGxtn7ohEROSMbNmuwz9+/DhOp5P9+/eza9cuvvzlL3Pf\nfffxyCOP8P777xMMBonFYtx66628/vrrJJNJbrzxRuLxODabDZ/Px+OPP47P56Oqqoof//jHVFZW\nzvTxiYjIZ7Ke0tm3bx9FRUUsXbqUzs5O6uvrAaivr2fPnj0AdHR0UFtbi91ux+12U1RURDQaJZVK\nMTw8jM/nA6Curs7sIyIisyMv2xe2tbVRW1sLQDqdxuFwAOBwOEin0wD09/dTXl5u9nG5XCSTSex2\nOy6Xy2x3Op0kk8nT3sNms53fUYiIWFi2GyZkNcLPZDL84Q9/4Lvf/e5pz9lstmkNasMw5uW/Bx98\nMOc1qP7c16H659+/+Vy7YUxuZ5ysAv/555/nmmuuYcmSJcDoqH5gYACAVCpFQUEBMDpy7+vrM/sl\nEglcLhdOp5NEIjGu3el0TqpQERGZmqwC/9lnnzWncwCqq6sJhUIAhEIh1q1bZ7a3tbWRyWTo7e0l\nHo/j8/koLCwkPz+faDSKYRi0traafUREZHaccw7/ww8/ZN++fTz55JNm2wMPPEBNTQ0tLS243W7a\n29sB8Hq91NTU4PV6ycvLo7m52ZzuaW5uZuPGjYyMjFBVVbXgVuj4/f5clzAlqj+3VH/uzOfaJyvr\nZZmzwWazTXpOSkTEyiaTm7rSVkTEIhT4IiIWocAXEbEIBb6IiEUo8EVELEKBLyJiEQp8ERGLUOCL\niFiEAl9ExCIU+CIiFqHAFxGxCAW+iIhFKPBFRCxCgS8iYhEKfBERi1Dgi4hYhAJfRMQiFPgiIhah\nwBcRsQgFvoiIRSjwRUQsQoEvImIRCnwREYs4Z+APDQ2xfv16li9fjtfrJRqNMjg4SCAQoLi4mIqK\nCoaGhszXNzU14fF4KCkpoaury2zv7u6mtLQUj8dDY2PjzByNiIic1TkDv7GxkaqqKt566y3efPNN\nSkpKCAaDBAIBenp6WL16NcFgEIBYLMbu3buJxWKEw2E2b96MYRgANDQ00NLSQjweJx6PEw6HZ/bI\nRERknLyJnjx8+DAvv/wyoVBo9MV5eSxevJjOzk5eeuklAOrr6/H7/QSDQTo6OqitrcVut+N2uykq\nKiIajXL55ZczPDyMz+cDoK6ujj179lBZWXnae27dutV87Pf78fv903SoIrMnP/9ShoffB2DRoks4\ncmQwxxXJQhGJRIhEIufVd8LA7+3tZcmSJdx+++38/e9/55prruGXv/wl6XQah8MBgMPhIJ1OA9Df\n3095ebnZ3+VykUwmsdvtuFwus93pdJJMJs/4nicHvsh8NRr2xmePbbktRhaUUwfC27Zty7rvhFM6\nR48e5cCBA2zevJkDBw7wxS9+0Zy+GWOz2bDZ9IEWEZnrJgx8l8uFy+XiuuuuA2D9+vUcOHCAwsJC\nBgYGAEilUhQUFACjI/e+vj6zfyKRwOVy4XQ6SSQS49qdTue0H4yIiJzdhIFfWFjI0qVL6enpAWDf\nvn2sWLGCtWvXmvP6oVCIdevWAVBdXU1bWxuZTIbe3l7i8Tg+n4/CwkLy8/OJRqMYhkFra6vZR0RE\nZseEc/gAu3bt4rbbbiOTybBs2TKefvppjh07Rk1NDS0tLbjdbtrb2wHwer3U1NTg9XrJy8ujubnZ\nnO5pbm5m48aNjIyMUFVVdcYTtiIiMnNsxti6yTnAZrMxh8oROW+jA52xz7I+1zJzJpObutJWRMQi\nFPgiIhahwBcRsQgFvoiIRSjwRUQsQoEvImIRCnwREYtQ4IuIWIQCX0TEIhT4IiIWocAXEbEIBb6I\niEUo8EVELEKBLzJN8vMv1R3gZE7T9sgi0+TULZG1PbLMBm2PLCIip1Hgi4hYhAJfRMQiFPgiIhah\nwBcRsQgFvoiIRSjwRUQs4pyB73a7WblyJWVlZfh8PgAGBwcJBAIUFxdTUVHB0NCQ+fqmpiY8Hg8l\nJSV0dXWZ7d3d3ZSWluLxeGhsbJyBQxERkYmcM/BtNhuRSISDBw+yf/9+AILBIIFAgJ6eHlavXk0w\nGAQgFouxe/duYrEY4XCYzZs3mxcENDQ00NLSQjweJx6PEw6HZ/CwRETkVFlN6Zx6FVdnZyf19fUA\n1NfXs2fPHgA6Ojqora3FbrfjdrspKioiGo2SSqUYHh42/0Koq6sz+4iIyOzIO9cLbDYbN954I5/7\n3Of4/ve/z1133UU6ncbhcADgcDhIp9MA9Pf3U15ebvZ1uVwkk0nsdjsul8tsdzqdJJPJM77f1q1b\nzcd+vx+/338+xyUisiBFIhEikch59T1n4L/66qtcdtllvPvuuwQCAUpKSsY9P92bRZ0c+CIiMt6p\nA+Ft27Zl3fecUzqXXXYZAEuWLOHmm29m//79OBwOBgYGAEilUhQUFACjI/e+vj6zbyKRwOVy4XQ6\nSSQS49qdTmfWRYqIyNRNGPgfffQRw8PDAHz44Yd0dXVRWlpKdXU1oVAIgFAoxLp16wCorq6mra2N\nTCZDb28v8Xgcn89HYWEh+fn5RKNRDMOgtbXV7CMiIrNjwimddDrNzTffDMDRo0e57bbbqKio4Npr\nr6WmpoaWlhbcbjft7e0AeL1eampq8Hq95OXl0dzcbE73NDc3s3HjRkZGRqiqqqKysnKGD01ERE6m\n/fBFpon2w5dc0H74IiJyGgW+iIhFKPBFRCxCgS8iYhEKfBERi1Dgi4hYhAJfRMQiFPgiIhahwBcR\nsQgFvoiIRSjwRUQsQoEvImIRCnwREYtQ4IuIWIQCX2TG5WGz2cjPvzTXhYjFaT98kWky0X74o4/1\n+Zbpp/3wRUTkNAp8ERGLUOCLiFiEAl9ExCIU+CIiFqHAFxGxCAW+iIhFZBX4x44do6ysjLVr1wIw\nODhIIBCguLiYiooKhoaGzNc2NTXh8XgoKSmhq6vLbO/u7qa0tBSPx0NjY+M0H4aIiJxLVoG/Y8cO\nvF7vZxeWQDAYJBAI0NPTw+rVqwkGgwDEYjF2795NLBYjHA6zefNm84KAhoYGWlpaiMfjxONxwuHw\nDB2SiIicyTkDP5FIsHfvXu68804zvDs7O6mvrwegvr6ePXv2ANDR0UFtbS12ux23201RURHRaJRU\nKsXw8DA+nw+Auro6s4+IFeXnX4rNZtOWCzKr8s71gnvuuYdHH32UI0eOmG3pdBqHwwGAw+EgnU4D\n0N/fT3l5ufk6l8tFMpnEbrfjcrnMdqfTSTKZPOP7bd261Xzs9/vx+/2TOiCR+WB4+H3Gtl4YHrbl\nthiZVyKRCJFI5Lz6Thj4zz33HAUFBZSVlZ31DcZGKdPl5MAXEZHxTh0Ib9u2Leu+Ewb+X/7yFzo7\nO9m7dy8ff/wxR44cYcOGDTgcDgYGBigsLCSVSlFQUACMjtz7+vrM/olEApfLhdPpJJFIjGt3Op1Z\nFykiIlM34Rz+9u3b6evro7e3l7a2Nr75zW/S2tpKdXU1oVAIgFAoxLp16wCorq6mra2NTCZDb28v\n8Xgcn89HYWEh+fn5RKNRDMOgtbXV7CNiHXnT/hexyGSccw7/ZGMf1AceeICamhpaWlpwu920t7cD\n4PV6qampwev1kpeXR3Nzs9mn
}
],
"prompt_number": 73
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"It is much easier to pick thresholds and compare the results of different template matches\n",
"in this kind of normalized histogram."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Computation via Convolution\n",
"===========================\n",
"\n",
"It turns out that the entire operation above can be expressed in terms\n",
"of the convolution. \n",
"\n",
"Consider the distance computation and expand the square:\n",
"\n",
"$d_{ij}^2 = \\sum (t_{uv} - I_{i+u,j+v})^2$\n",
"$ = \\sum t_{uv}^2 - 2 \\sum t_{uv} I_{i+u,j+v} + I_{i+u,j+v}^2 $\n",
"\n",
"Now it is easy to see that the individual terms can be written as\n",
"follows:\n",
"\n",
"$ d = ||t||^2 - 2 I * \\hat{t} + I * b $\n",
"\n",
"Here, $\\hat{t}$ is the flipped version of the template $t$, and\n",
"$b$ is a box of the same size as the template.\n",
"\n",
"The operation $I * \\hat{t}$ is called the *cross correlation* between\n",
"the template $t$ and the image $I$."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"t2 = sum(template_A**2)\n",
"cc = filters.convolve(textimage,template_A[::-1,::-1])\n",
"I2 = filters.convolve(textimage,ones(template_A.shape))\n",
"out = t2-2*cc+I2\n",
"imrow(ds,out)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAACGCAYAAAA8c1eJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nuazxeKzxeBwygfyRK43cMv/I
}
],
"prompt_number": 27
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"As you can see, up to boundary conditions, the results are the same.\n",
"\n",
"Writing the template matching operation in terms of the convolution lets\n",
"us apply all the computational tricks (in particular, the Fourier transform)\n",
"to this problem."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(cross correlation for template matching)\n",
"\n",
"Many people use cross correlation as if it were a \"template matcher\"\n",
"all by itself. It really is not. \n",
"\n",
"Cross correlation for detection of signals usually assumes zero signals \n",
"and templates with an (expected) zero mean. \n",
"\n",
"In that case, $||t||^2$ and $I*b$ are expected to be close\n",
"to zero. If you really know that that is the case for your signals, then\n",
"you don't have to bother computing $||t||^2$ and $I*b$. \n",
"\n",
"But usually, computing all terms is a better choice."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Normalized Cross Correlation\n",
"========="
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Above, we derived template matching and related it to the convolution (and cross correlation)\n",
"by considering the Euclidean distance between a template and the image.\n",
"\n",
"There is a second way in which we often determine whether two vectors are similar:\n",
"we determine whether they are colinear, or, more generally, what the angle between\n",
"them is. This is done easily by the dot product:\n",
"\n",
"$\\cos(\\angle(u,v)) = \\frac{1}{||u||} u \\cdot\\frac{1}{||v||} v $\n",
"\n",
"As before, we compute the norm of the template directly, but\n",
"we compute the norm of all the image patches by convolution,\n",
"and the dot product between the template and the image by\n",
"convolution as well."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"tl = sqrt(sum(template_A**2))\n",
"cc = filters.convolve(textimage,template_A[::-1,::-1])\n",
"Il = sqrt(filters.convolve(textimage**2,ones(template_A.shape)))\n",
"cout = cc/maximum(1e-5,tl*Il)\n",
"imrow(ds,cout,cout**8)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABBAAAACDCAYAAAA02OgOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzsnUuPZNlV/Ve8MuP9yEdlVlW3acs0buNuhGwe4jEAhGU8\nseQRzBgY5AkDRvAVbIkhnwGrmcEEDzzAAga0PUCWjZEtN253d1Vl5SsiMt7P/yD127HiVFR38XeL\nxl33SKmsiox77zn77LP32mvvc25uvV6vlbWsZS1rWcta1rKWtaxlLWtZy1rWsvYeLf9hdyBrWcta\n1rKWtaxlLWtZy1rWspa1rP3fbxmBkLWsZS1rWcta1rKWtaxlLWtZy1rW3rdlBELWspa1rGUta1nL\nWtaylrWsZS1rWXvflhEIWcta1rKWtaxlLWtZy1rWspa1rGXtfVtGIGQta1nLWtaylrWsZS1rWcta\n1rKWtfdtGYGQtaxlLWtZy1rWspa1rGUta1nLWtbet2UEQtaylrWsfQTbN77xDb3yyit6+eWX9bWv\nfe3D7k7Wspa1rD13LbPDWcta1j6KLbder9cfdieylrWsZS1rH1xbLpf65Cc/qW9+85u6f/++fv3X\nf11f//rX9alPferD7lrWspa1rD0XLbPDWcta1j6qLatAyFrWspa1j1h744039Iu/+It66aWXVCqV\n9Cd/8if6h3/4hw+7W1nLWtay9ty0zA5nLWtZ+6i24ofdgaxlLWtZy9oH29599129+OKL8f8XXnhB\n//7v/771nVwu97/draxlLWtZe6b2USiOzexw1rKWtZ/39jRbnBEIWcta1rL2EWvPCkp/53d+R3t7\ne1qv1zo6OlKpVNJgMNB6vdb+/r7u3LmjdrutfD6vbrerBw8eaD6fq1KpqFwuq9PpqFAoqNvtajwe\na29vT/v7+6pUKmq1WhqNRjo7O9NwOFS5XNbh4aEajYYkaTgc6vz8XN1uV61WSycnJyoWi7q6utJi\nsdBisVCxWNTe3p6KxaKKxaKWy6WGw6Gm06mWy6UajYYqlYpms5l6vZ6Wy6XefPNN/cZv/Iaq1aqK\nxaLm87murq40n8+1t7encrms09NT5XI5nZ+f6/Hjx1oul2q1Wjo4OND+/r5Wq5X6/b7Ozs40mUx0\n584dHR0dablcajKZ6OrqSrPZTJVKRc1mU+VyWZI0Ho81m800mUwkSZ1OR/v7+5rP5xqNRloul1qt\nVjo4ONDR0ZGq1WrMVz6fV7FY1OHhocrlsnq9nt555x1J0re+9S39/u//vsbjsfr9viaTiSaTicbj\nsVarlWq1miqVinK5nCaTibrdrtbrtdrtthqNhkqlkpbLpebzuS4uLmIOT09PY56Gw6HOzs40GAxC\nFu12O+ZqMpno5uZmS3/y+bzy+bz29/dVKpW0v7+vXC6n+Xyu8Xisy8tL5XI5tdvt+NtisdBsNtO3\nv/1tvfrqq2o0GiGnarWqfD6v6XSq/f19feITn1CpVNJ6vY7+XVxcqFwuq1AoaDqdajAYaDqdqtvt\najKZqFQqqVarqVwuazab6ebmRuv1WsViUeVyWbVaTfl8Xjc3NxoOhzGOSqWier2u+Xyum5sbzWYz\n7e3thU6USiWNx2P1er2t5+Tzt4Wcy+VSo9FI3/nOd/Sbv/mbarfbKpfLWq1WGo1Gur6+1s3NjfL5\nvNrttqrVqhaLhabTqYbDoQqFQshgb29Pk8lE19fXmkwmWq/X6nQ6Ojw81Gw209XVVehYrVZTq9VS\nsVjUeDxWt9vV1dVVrI+Pf/zjarVamk6neuutt3R+fq5isahOpxOyZ/0MBgNJUqvVinUwmUz0xhtv\n6BOf+ISq1WroPH8fDAa6ubnRarUKvSiVSioUCiqXyyqXy9rf31ehUNDe3l7Yk2KxqHq9rlqtpu9/\n//v63ve+F9e//vrrz2rq/k+3Z7XD1WpVzWYz9HS1Wmk2m2m9XqtQKKhWq6lWqymXy2k0Gqnf72u1\nWimfz2tvby/W/nQ61Xw+V6FQUC6XC1vMOlgsFioUCqpWqyqVSioWi5pMJhoOh5rNZiqVSmo0GvE5\nfcjlcioUCsrn8yoUClqtVlqtVppOp1qtVtrb2wsbs1wuNZvNNB6PdXh4qL29PeXz+bA/0+lUuVxO\ne3t7ajabyufz6vf7Gg6HWq1W0b9qtRr6NxwOwx/UajXt7+9rOp1qPB5ruVyGf+CHvi0Wi+g7Nhrf\nslqtVK1W1Wg0oo+SVCgUVCwWdXR0pEqlouVyqZ/+9KcaDod6/PixXnzxRQ2Hw3g296N/pVJJ+Xxe\ni8UiZFEqlVQul1UqlaIPLotyuaxWq6V8Pq/BYKDBYKDFYiFJqlQqqtVq2tvb02Kx0Gg0iv6n+pXP\n5+P5yGGxWGgymWg+n8ecF4tFrdfr6H+/3w8bwhrn+uVyqaOjI927dy9s++Xlpc7OzrRcLrW3t7f1\nnMlkoul0Kknhv3O5XMiCfu7v76tYLIacuAY7mM/nNZvNQqfQC7dLs9ksdIZxo5uXl5cht3K5HPKj\nj8vlMvqxt7cXfeB+yBIMgQxZj/R9MplosVgon8/HcySFD2Rc9Xpd7XZbhUJBs9ksMBPzX6/X457T\n6TSuK5fLgX2ur69VqVS0WCyib/wdHZxOp1sBP+PI5XIxH25/uQd9ePz4sc7OzuIe3/3ud59qtzIC\nIWtZy1rWPmLt/v37evvtt+P/b7/9tl544YUnvveZz3xG7XZbk8lEg8FA3W5XkgJ4zWYzzefzAKcE\nM4AsAkKc73K5DHCYz+fjHgQm0i2bDXjBmddqNZVKJZVKpSA0FotF/J7NZuEoAQkAx1qtJkkBUKVb\n4L5erwNg4MwlxT0AjKPRSJPJRJVKRZLi+wBUwAYOG8c6mUzi/wA6v65SqYTjps/T6VSj0UiSAogX\nCgVJCqfearWiH6PRSIVCIeQ1nU51c3MTRASgwQkEPlsulyFr5MA14/E4ngtYkhT3L5VKqlQqqlar\nAbzn87lms1kAUNeTcrkc9xiPxwGCu91ugC2A3HK5jIBzOp0GsHMASKBD/2mLxUK9Xi/GjEzQo8Fg\noGq1qlarFaDQgStyy+fzmkwmGo1GEYS5HNBNSATGPRwOg1yo1+tBLACsITOurq7iXgQTgDvkDtAk\nmEDP0C2CyPl8rvV6rdVqFQGIpJhTiBnWI9cDdn1d5nK5CPTR29FoFAHfYDAIPa/X67FeWCfoIUSh\ny4v1uF6vYx0y/vV6Hc8BsLstePXVV/Xqq6/GNR8VAuFZ7fD+/r6Ojo7CfkAOMe8Eem5v+TtkAd9z\nHUevCeDQJRrf5/7FYjFsO4ETgSr3X6/XYduZP4I+dA0d4ztui+h7aoexofTP7RVjxpZCVjAG7Ab3\n8eAYu4K95TvIE6LFyU8CMtbTdDqN9eZEiNsn5gqijL7vGpPLAhm5LOg7fYWE4DlPmwf6C2EDoURg\njCx8rNyLviBXJ0HcZvAbmfBd7Ak/6AX2znWZfqdz5USVj4s+IHMfGw3SQ1LgFObMiRZfJ8gMuSN7\n1o//P/0c/cVWuy74/Zjz+XweY+TvPN9lh/2cz+ehE2AiZMkcoW+M2+XMM3g+ckUvSais1+vQV0k6\nPT3V6elp3O9/lUD4xje+ob/8y7/UcrnUn/3Zn+mv//qvP+hHZC1rWcta1t6j/dqv/Zp+9KMf6Sc/\n+Ynu3bun119/XV//+tef+B4AQLrNGlIhQKaZoC+Xy0UWZX9/X6enp6pUKhFc4dSXy6UqlYoajYba\n7Xaw24Cqer2+lXUn69Zut9XpdOJvBO44Rqob6vX6VqB5enqqO3fuRPXBZDLR+fn5FtOfz+cj2wvL\n3mw2VSwWNRqNAox6FoCxDAYDlctltdtttdttlUoljUYjjUYjrVaryCqQ+cdZr1YrdTod3b9/X6VS\nKYJOQAIBG4EY4ISMtbQJrgCSDoI8YzCdTnV8fKzT01MVi0VdXl4G8CPTTSAg3WY7madmsxmZJ+67\nt7cXWSPmDxCWz+cju8+YZ7NZ
}
],
"prompt_number": 28
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Now, there is one more improvement we can make to this.\n",
"As written above, if we replace $u$ by $u+w$ and $v$ by $v+w$ for\n",
"$w>>u$ and $w>>v$, then the new vectors will be nearly colinear no\n",
"matter what $u$ and $v$ are. \n",
"\n",
"For cross correlation to work well\n",
"in detecting a pattern, we want $u$ and $v$ to have a wide variety\n",
"of directions, and to accomplish that, we subtract their means."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"template_A_centered = template_A-mean(template_A)\n",
"textimage_centered = textimage-filters.convolve(textimage,ones(template_A.shape)/sum(ones(template_A.shape)))\n",
"imrow(template_A_centered,textimage_centered)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAArsAAAFKCAYAAAAdRZejAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n+q7RaITJZEIymVTpm5qaUFZW
}
],
"prompt_number": 29
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"tl = sqrt(sum(template_A_centered**2))\n",
"cc = filters.convolve(textimage_centered,template_A_centered[::-1,::-1])\n",
"Il = sqrt(filters.convolve(textimage_centered**2,ones(template_A_centered.shape)))\n",
"cout = cc/maximum(1.0,tl*Il)\n",
"imrow(ds,cout,cout**2)\n",
"print amin(cout),amax(cout)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"-0.457507 0.950109\n"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABBAAAACDCAYAAAA02OgOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nunv3rsbjsXq9nubzuXq93gYQ
}
],
"prompt_number": 31
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"There is actually yet another way in which we can express this.\n",
"Instead of computing the cross correlation in one step like above, we can\n",
"think of this as \"standardizing\" the image, filtering it with a nonlinear\n",
"filter that achieves a constant mean and variance across the image.\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"textimage_standardized = textimage_centered/maximum(1e-5,Il)\n",
"imrow(textimage,textimage_centered,textimage_standardized)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABBAAAACDCAYAAAA02OgOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n47ON/ltPPNCfaX+TpMGCH5uj
}
],
"prompt_number": 32
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can do the same to the template, but of course \n",
"its display doesn't change (since this amounts to just a linear transformation \n",
"of the gray values, which the display routines compensate for).\n"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"template_A_standardized = template_A_centered / sqrt(sum(template_A_centered**2))\n",
"imrow(template_A_standardized,s=2)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAI0AAACMCAYAAABBAioAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nXNKV53a3sqpuJcmc1nT4r5X7
}
],
"prompt_number": 33
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"In terms of these standardized images, the normalized cross correlation now just requires a convolution with a flipped template.\n"
]
},
{
"cell_type": "code",
"collapsed": true,
"input": [
"normalized_cross_correlation = filters.convolve(textimage_standardized,template_A_standardized[::-1,::-1])"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 34
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(textimage,normalized_cross_correlation,normalized_cross_correlation**2)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABBAAAACDCAYAAAA02OgOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nVq6nPigWiwgGg2hsbEQsFpN9
}
],
"prompt_number": 35
},
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Peak Finding"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Many computer vision algorithms involve the following steps:\n",
"\n",
"- remove noise\n",
"- measure similarity to some template\n",
"- find a small number of locations where the template is present\n",
"\n",
"We find locations by identifying places of maximum similarity to the template."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Issues in peak finding:\n",
"\n",
"- scores near an actual instance are also high\n",
"- global maximum gives us only one location when there may be multiple instances\n",
"- local maxima may give us locations with low absolute score\n",
"- noise produces false maxima\n",
"- we often don't know what the threshold for a good score should be"
]
},
{
"cell_type": "code",
"collapsed": true,
"input": [
"ncc = normalized_cross_correlation"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 36
},
{
"cell_type": "code",
"collapsed": true,
"input": [
"def local_max(image,sigma=1.0):\n",
" image = filters.gaussian_filter(image,sigma)\n",
" return ((image>roll(image,1,axis=0))*\n",
" (image>roll(image,-1,axis=0))*\n",
" (image>roll(image,1,axis=1))*\n",
" (image>roll(image,-1,axis=1)))"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 37
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's start by finding local maxima in the ncc output. We can display either their locations,\n",
"or their locations with their weights."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(local_max(ncc),local_max(ncc)*ncc,s=10)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABH8AAADLCAYAAAD3Pz/yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3X1sVtd9wPHfA3YgYF4saoxnWxjZxjbg+UXmVSEFgkNg\ngkC6MMwakZBI1ar90W7a2v4xbdW0hmyatDdVqjplQp3aZJUWIFVKSZbQpCWNw4AsCZQCNZntGESI\nMWACxubuD4aDjf089+W83Xu/HylS8HPvOb/n3Pvc55zfc+85Gc/zPAEAAAAAAEAiTbAdAAAAAAAA\nAPQh+QMAAAAAAJBgJH8AAAAAAAASjOQPAAAAAABAgpH8AQAAAAAASDCSPwAAAAAAAAmmPPmzf/9+\nqa2tlerqannuuedUFw8AAIAx0AcDAADjyXie56kqbGhoSGpqauS1116T0tJSWbx4sfzoRz+Suro6\nVVUAAABgFPpgAAAgmzyVhbW3t0tVVZVUVFSIiMi2bdtk7969wx2PTCajsjoAAOAohb8twQf6YAAA\nIFv/S2nyp7u7W8rLy4f/XVZWJu+8847KKgAAADCKnz7YpEmTZPLkyaZDS6Xr16/T1gbR3ubQ1mbR\n3uYkoa37+vqyvq50zh+TvyqF+UUxbr9C6oo3bu1wR1zj1oX2SBaO572Scg2M07GNU6wYyWQfbNGi\nRYH3KSkp0RCJPvPmzYtVubrNnTvXdghOietxxL2qq6tth+AcXef33T9QmBCnz2lRUZGRepQmf0pL\nS6Wzs3P4352dnVJWVpZ1n7AdzTCdHL/7BI3J7/ZBy9XVkQtarq72CEp3xzZug54w7RH2Papsm7i1\ns0j0mP3sr/N46mxzneeUrjYJey0x+X1li+7vB+gTpg82Z86cUHV98MEHgffp6enxtV3QDvCdx9xy\nKSwsDFRuR0dHoO11lRs0aaZrcPXRRx9pKfeOmTNnai1ftTDnR67P41hUJyZKS0uVlmdClEHxtGnT\ncm5z6tSpwOVOnTrV13ZhjrlfYRMbfmIPc377ucbe/R0VRNi563Rdx0XUfzYvXLgQaPuw10ylEz4P\nDg5KTU2N/Nd//Zf8zu/8jixZsmTEZINx6gAjnTzP4zwFAAVIDJnlpw82depUyctT+sQ/xjE4OEhb\nBzRjxoycjyyMh/Y2h7Y2i/Y2Jwlt3dfXZ27On7y8PPmXf/kXWbdunQwNDcnTTz/NKhOIFRI/yUNC\nD0Aa+OmDxb1TGye0dXBhEz8itLdJftt61qxZcvHiRc3RJB/ntjlpaGuld/7krIwBGBLATzKBhAOA\ntOPOH7dkMhmZMWOG7TCASCZNmiQ3btwY9/WSkhLfjxgCQNLkuvNH6Zw/QBr4SeqQ+AFgAgkWAGmS\nLfEj4n9uKQBIo9gmf+La4dU1OWpc20Mk3rHr4Gp7RIlLxXtytV3wOReOke4YXJs8O2qi2YVjhvhp\nbm62HUIoq1atCrVfTU1N1tfXrVsXqlwXhG2TpNq4caPtEO7R0tISaf8lS5YoigSu0jmxtF8PPvig\n1vJXrFihrWwbn3tb36NOJn90rtaiqv6w/DwuFGa/uLaHiL7YXR7UZIvNT3vYWLUr7LnrZ18V9Zvi\nyup3LiY5ghwjF1YFtJFssVH2eO+Tx1Mhcu8S3n5WkDly5IiucGTZsmXayj548GCo/U6ePJn19Z/9\n7Gc5ywg7B+YDDzwQaj+/wrZJLuvXr9dSblS5Vuh5+eWXc5ZRW1sbqu41a9aE2u/w4cM5t6msrBz3\ntfb29lD1uiboSna6BvStra1ayr3jqaeeCrxPV1eX7203bNgQuHw/3nzzTd/bhknkHDp0KPA+fvn5\n3Icx3vWmqalJ6/doNqmY8yduHdy4xZs0cWn/uMQJIJ1cTsCnka05f5YvXy5vv/228XrDYL4Y+woL\nC6W3t9d2GDnV1NTkTAoCgGnM+SO3Ozxx+vVd94DetccWbJY7lrgkVHSe1wDU4DOKtHv77bdl+fLl\nWsoOeyfFeEwkfnI9QhbF2rVrtZTb1NSkpdyxxCHxI3L7brBcdxEBsIcVx8eWijt/AKQHd0QlC8cz\nvkh8uYXVvgDoNmfOHDl37pztMKDAokWL5IMPPrAdBgLizh8AqWIiUZCEO9x0iOOdkAAAQA3diR9d\nd1s1NDRoKdeUqqoq5WWS+Ekm7vwBhLsLkFyc27Al7snMpOHOH7iKu0WQROXl5dLZ2Wk7DKRMYu/8\noVMJlRgcI6k4t9Xhewe4zc+KYIBfJH6QRCR+1NE1b1wacecPAABQjmSZW7jzBwCAZEvsnT9hpKkj\nmqb3aprptnX1WLoaFxBWEs5pv++Beatg2ooVK2yHYMzEiRNth5BY+fn5tkNwAucYkmZoaMh2CJEU\nFBT42q6iokJL/aWlpb62S1XyJ013HqXpvZpmum1dPZYm42JACRNsftZUneN+34Ou9+rq9Qr2HTp0\nyHYIxsR9EOOymzdv2g7BCSbPsZ07dxqrC+llK6H5xBNPKCnn6tWrvrY7e/askvpG6+7u9rWdM499\nMSkpACDNkvY9SNLWLdke+2psbJRjx44ZjggAAPva2trkRz/6ke0wlMj12JczyR8AAJAcJH/cwpw/\nAAAkW6zm/HGlo6gzDlfeYxBxbo84trdOtIdaNudXMXUs4/b5j1u8gCsaGhpshyAiIl/+8pe1lb15\n82ZtZeui85GbLVu2aCtbROQv//IvtZYfN88884ztEBKjpqbG13Y6zvGvfe1ryssci85VFTdu3Ki8\nzMcee0x5mXekac443ZxK/vi5Myhs5zvIfkHvUAoyAFRx91OUAWeY9tPVHmHKNlG+qrpVlmdyPhAV\nddkYJNuoM5PJ+KrX73mo49yO2i5BPkNB69JxPtqON2zZLpZFsitd3nvvvZzbtLa2hio7yKDg3//9\n3wOV3dzc7DuGPXv2BCp7LGvWrPG13VhxzZ07N3B9zz//fKDt6+vrfW/70ksvBQ1HZs2a5Xvbb3/7\n24HLz8bvhKp+rFu3ztd2S5YsUVbnv/7rv+bcpri4OFIda9eujbR/WH4/h6qcPHlSamtrc27n9xyv\nrq72Xfc//MM/+NrOb4JqPB0dHb63ffDBBwOV/fLLL+fcJuj16j//8z99b1teXh6o7CBzxjU2NgYq\nO5uFCxcqK6utrc33tgMDA8rqHY3HvgBYlbR5TgDcRvLILTz2BeBu/f39MnXqVNthAFAoVo99AUgf\nEj8AAKRLYWGh7RBSj8QPkD4kfwAAAAAY09vbazsEAEgdkj8AgMh4xAcAAMCsqHMLIV2cS/4wgACA\n+OHxvWTiOzldioqKbIcAAAjg5MmTtkOABv39/VrKdS75wwACfjAgAWBKmOtNUq5RfCeny4ULF2yH\ngBjQuQQ1gHul9bv4i1/8YuB9knJ90jUnl3PJnyQzMRiI04AjSqxJuwiO1Ra2jmWUenXFHKfz+o6o\nMcfxPfsRx3MkzPUmzD5JPeaAC3bu3Km9jpUrV2qvQ5Unn3wy9L5BlqCOg0ceeeSev4UZdKoQdal3\nHTZs2GA7BKNcnIxcRf9A13F85plntJQrIvLzn/888D5hrk/r1q0LvE9csdQ7gNRjuXlAPZJZbmGp\ndwCumTlzply6dMl2GEBisNQ7AORA4gcAAMAsEj+AWSR/gAj4ZRsAAMA87mQDgGBI/gARcMcIAACA\neX19fbZDAIBYIfkDAAAAAACQYCR/AMQej98BAACY5+IKZQDGRvIHQOxFffyO5BFU4nwCAKTF+fPn\nQ+/LvE1QacqUKbZDcB7JHwBKxHnAq3rupji3RdxEbWsdx8rVucA4L5EUN27csB0C4JS6ujrbIYSi\nY96mlpYW5WXiXlHv+NLRJ7l27ZryMlWor6+3HcIwkj8JQsfeDtXtHtfj6OqA1wbawgzP8yK3dZqO\nVZreK5Jt0qRJtkO4h83O/dWrV63Vbdu0adOUltfc3Ky0PL/y8vIi7X/ixAlFkcTf4cOHbYeQeMXF\nxZHu+BJJV5/k/ffftx3CsIxn
}
],
"prompt_number": 242
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's threshold these (since there are too many). We take, as a threshold, 0.8 of the peak value."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"gm = amax(ncc)\n",
"lm = local_max(ncc)*ncc*(ncc>0.8*gm)\n",
"lm /= amax(lm)\n",
"imrow(lm,s=10)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAlEAAADdCAYAAACbkSdUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nutVqZX5+vlPD8wELCwtpNptJ
}
],
"prompt_number": 250
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's overlay this on the original image, spreading the values around \n",
"as much as the size of the original template."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"lm30 = filters.maximum_filter(lm,30)\n",
"imshow(transpose(array([textimage,lm30,zeros(textimage.shape)]),[1,2,0]))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 251,
"text": [
"<matplotlib.image.AxesImage at 0x4521290>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAACQCAYAAAAV3+A0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nv0AoxxJCIZ+igscQAWp/QSjg
}
],
"prompt_number": 251
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"It is possible that we get multiple peaks per character.\n",
"To avoid this, we can use a simple non-linear filtering operation."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"lms = lm*(lm>=filters.maximum_filter(lm,30))\n",
"lms30 = filters.maximum_filter(lms,30)\n",
"imshow(transpose(array([textimage,lms30,zeros(textimage.shape)]),[1,2,0]))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 247,
"text": [
"<matplotlib.image.AxesImage at 0xae56a10>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAACQCAYAAAAV3+A0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nv0AoxxJCIZ+igscQAWp/QSjg
}
],
"prompt_number": 247
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Less obviously, another good way of getting rid of multiple\n",
"peaks is with Gaussian filtering."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"lmg = filters.gaussian_filter(lm,15.0)\n",
"lmgs = local_max(lmg)\n",
"imrow(lmg,lmgs,s=8)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAA6AAAACpCAYAAAA1IlDkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nASDNpqYAALa2uQugTVb7y7s9
}
],
"prompt_number": 239
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"ncclms = where(lm<lm30,0,lm)\n",
"lms30 = filters.maximum_filter(lms,30)/amax(lms)\n",
"imshow(transpose(array([textimage,lmg30,zeros(textimage.shape)]),[1,2,0]))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 201,
"text": [
"<matplotlib.image.AxesImage at 0x4018890>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAXIAAACQCAYAAAAV3+A0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhPs5QjmWEQr5OBU8RghQ+wtC
}
],
"prompt_number": 201
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's see how this works by looking at a bunch of points in well-defined locations,\n",
"and seeing them merge for increasingly larger sigma."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"test = zeros((30,100))\n",
"test[15,5] = 1\n",
"test[15,10] = 1\n",
"test[15,20] = 1\n",
"test[15,40] = 1\n",
"test[15,80] = 1\n",
"imshow(test)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 213,
"text": [
"<matplotlib.image.AxesImage at 0x6ecb890>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAACBCAYAAADkMoBRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ne4QQQpZjMTExIXJycu67X45j
}
],
"prompt_number": 213
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"testg = filters.gaussian_filter(test,3.0)\n",
"imshow(testg)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 206,
"text": [
"<matplotlib.image.AxesImage at 0x9fac510>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAABfCAYAAADMBv0OAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n18qwXTXavIaJcmB5sGOGAjEH
}
],
"prompt_number": 206
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imshow(local_max(filters.gaussian_filter(test,3.0,mode='constant')))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 223,
"text": [
"<matplotlib.image.AxesImage at 0x7b88050>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAACBCAYAAADkMoBRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n82L79u3i3LlzCTkWqxGVK2y7
}
],
"prompt_number": 223
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imshow(local_max(filters.gaussian_filter(test,1.0,mode='constant')))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 222,
"text": [
"<matplotlib.image.AxesImage at 0x79ae850>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAACBCAYAAADkMoBRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ne4QQQpZjMTExIXJycu67X45j
}
],
"prompt_number": 222
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imshow(local_max(filters.gaussian_filter(test,5.0,mode='constant')))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 221,
"text": [
"<matplotlib.image.AxesImage at 0x9942650>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAACBCAYAAADkMoBRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTZsmt9v9jx/3TJaxGB4eVmNj
}
],
"prompt_number": 221
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imshow(local_max(filters.gaussian_filter(test,10.0,mode='constant')))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 220,
"text": [
"<matplotlib.image.AxesImage at 0x97e0290>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAACBCAYAAADkMoBRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nuHGjWlpalJWVlXJj8eOPP+rp
}
],
"prompt_number": 220
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"General strategies for peak finding:\n",
"\n",
"- if you don't know a good threshold, use a fraction of the global maximum\n",
"- you also should have a conservative absolute threshold\n",
"- use local maxima after (and only after) Gaussian smoothing\n",
"- Gaussian smoothing will also ensure separation of locations (but make localization less accurate)\n",
"- you can complement Gaussian smoothing by inhibiting local maxima close to other local maxima"
]
},
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Peak Finding with Mathematical Morphology"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"def showmax(a):\n",
" \"just broadens maxima a little with a maximum_filter for better visibility\"\n",
" return filters.maximum_filter(a,3)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 52
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(showmax(local_max(ncc)),showmax(local_max(ncc)*ncc),s=10)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABH8AAADLCAYAAAD3Pz/yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nh8/8MQuvs2SCJCZ/iIii4CZ/
}
],
"prompt_number": 53
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"peaks50 = (ncc==filters.maximum_filter(ncc,50))\n",
"imrow(showmax(peaks50),showmax(peaks50*ncc),s=10)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABH8AAADLCAYAAAD3Pz/yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n1tbW0m3f/OY3o7GxMdrb26O9
}
],
"prompt_number": 54
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"thresh = (ncc>amax(ncc)*0.7)\n",
"imrow(showmax(thresh),showmax(peaks50*thresh),s=10)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABH8AAADLCAYAAAD3Pz/yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAqpvwBAAAAqJjyBwAAAKBi
}
],
"prompt_number": 55
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"unravel_index(argmax(ncc),ncc.shape)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 57,
"text": [
"(166, 194)"
]
}
],
"prompt_number": 57
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"plot(ncc[166,184:204])"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 64,
"text": [
"[<matplotlib.lines.Line2D at 0x3805fd0>]"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD9CAYAAABQvqc9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nwxX4p3NiikdjJakJjfTF53zz
}
],
"prompt_number": 64
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"ys,xs = mgrid[-10:10,-10:10]\n",
"r = hypot(ys,xs)\n",
"imshow(r)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 63,
"text": [
"<matplotlib.image.AxesImage at 0x38176d0>"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAPwAAAD5CAYAAAADZljUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n+lrgS5/QE03wKb1l88psdtiz
}
],
"prompt_number": 63
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"upper = -0.1*r\n",
"lower = where(r<1.0,0.0,-2.0)\n",
"plot(ncc[166,184:204])\n",
"plot(upper[10,:]+ncc[166,194])\n",
"plot(lower[10,:]+ncc[166,194])"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 72,
"text": [
"[<matplotlib.lines.Line2D at 0x49b9050>]"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD9CAYAAABQvqc9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nX8glqzNn9H9e6Rz/5L7JiNsV
}
],
"prompt_number": 72
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from scipy.ndimage import morphology"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 73
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"ncc_u = morphology.grey_erosion(ncc,structure=upper)\n",
"ncc_l = morphology.grey_dilation(ncc,structure=lower)\n",
"mpeaks = -(ncc_u-ncc_l)\n",
"mpeaks50 = (mpeaks>0.5*amax(mpeaks))*(mpeaks>=filters.maximum_filter(mpeaks,50))\n",
"imrow(mpeaks,mpeaks50,s=10)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABH8AAADLCAYAAAD3Pz/yAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3UuMLOdZN/B/3S/dPZdzGzvn+BJiJybCgaAQNkQgBfPt\nrGQTEpBikbCBJazYkQWKWYOQEBKSVyRsSLLBilgEISJhhAggAnGU2LFz7nNmuqdvda9vcXjeU93T\n3dOX6u6qnv9PGvl4zkx3ddVbdep56nmfV8vzPAcREREREREREe0kfdsbQERERERERERE68PkDxER\nERERERHRDmPyh4iIiIiIiIhohzH5Q0RERERERES0w5j8ISIiIiIiIiLaYUz+EBERERERERHtsNKT\nP2+++SZeeuklvPjii/jTP/3Tsl+eiIiIiCbgPRgRERFNo+V5npf1Ymma4iMf+Qj+4R/+ATdv3sQv\n/dIv4W/+5m/wsz/7s2W9BRERERGN4T0YERERzVJq5c9bb72FF154Ac8//zwsy8LnP/95fPOb3yzz\nLYiIiIhoDO/BiIiIaBazzBe7ffs2nnnmGfX/t27dwr/8y7+M/IymaWW+JREREVVUicXFdAHegxER\nEREw/f6r1Mof3lQQES1H13VYlgXTLDUnvxP4b8t0tm3DMIxtbwZVAM8TIiIimqXUKOPmzZt4//33\n1f+///77uHXrVplvQUS0k7IsQ5Zl296MSmL1yHRRFG17E6gieA9GREREs5Ra+fOJT3wCP/zhD/Hu\nu+8iiiJ8/etfx6uvvlrmWxARERHRGN6DERER0SylVv6Ypok///M/x//7f/8PaZriy1/+MleZICIi\nIloz3oMRERHRLKUu9T7XG3JOOhER0aXAKXvVwnswIiKi3beRhs+0HcWbOV3nIa3rza1pmmrb2fi3\n/uR4Fr/qwrIsWJa17c0gIiIiIqKSsPKHFqLrOpvSEi1Brn1VroSQ5LGmacjznOf6/5H9sQ6GYcCy\nLCRJgiRJ1vIe21Tl8X4Z8R6MiIho97HyZ0mspHnCMAzujw0pcz9rmsbjVgHNZhO+71c6+LIsC57n\nqcSP4zhrrUCzLAuu667t9esgz3PkeV7pcUFERERE9bfViLDqAallWVO38TJOybEsC4ZhTPy7y7g/\n1qnMigtd10emlNF25HmONE0rXwkhycJNJA0l8XGZZVmGOI4Rx/G2N4WIiIiIdthWsy+GYWwlIHVd\nd2qyQr5vmiYcx5mY7NA0Db7v4+DgYK3buW22bavgz3EcOI4ztQ+Ibdu4cuUK+4RUkKZpcBwHnuep\n70lwT5sTxzHSNN32ZsyUpqmafpTn+dq3eVenOi2K0+uIiIiIaN22lvzRNA2GYcC27Y2+b6vVgu/7\ncBznXPBrGAZ83wfwOEE0bcqDZVmwbRue5218+9fBNM2JT/ht21b7Q5I/kxJ2mqbBNE34vo9ms7mR\nbabJLMs6d3x0XVfjVRoPT/o5Wq8syyZWuVSpGfT4duxiUqIq+5qIiIiIaJO2Pu1rk8kTeT/TNKdO\nN5Ckj1S9TPqZKIpUif6uBBKTgrxikmDW/pAKAfaW2S7TNCcmGGSMmqYJ27ZVNduujN26mJbkMU0T\njUZj6pTKTZk0PXDdialpFWhyzSn7vSVRvajx6xqnuRIRERFR3Ww9Ut9kAsj3ffVek6YySPAjy2zP\nmhoTRdFOPBXXNG3qtA7DMGCaJkzTvDAwzbIMaZruxD6ps0nJOV3X1bGUJbzZA2jzpl1PkiRBlmVb\nT5zKGBnfjnVPEZz0uaU/Utnvu8wKZpLULk6B3YWKTyIiIiK6XC5V8kemaU17mq3rOlzXVT83bToU\n8CSI2OVkhyxBLMGOJAwm7TuppFplfzAZsR5S7SDH8qLEJq3HtH2e5zmiKFq6t05Zx1GSvMXXk8Th\nuqqSZr32upLJi+5nubZJYrWYCFqVVFduO/FHRERERLtva3eccjOt6zocx1nre0kw47quutGeVCEh\nPYgk2SGrW80Kri5D8qdYLTJrf8jT+mVJgqlM0lvqMiv215IxXfXkj1wXJjUQ31aj+DKsY7+XkTjQ\ndV2d58XXk6mC60pQrPP8lKrFMhSnU5a5Opnrumi1Wmi1WiPfZzKIiIiIiMq21cYFeZ4v3YNhEZZl\nIU1TFfhOq1DRdR2+7yPPc1iWhSRJ1O+Mr0iT5zmSJKn86j0XmRXI2LYN13WRJAl830eSJAiCAIZh\nnPvc0vcnDMOltkPeSyogygqwJOE3HA53finlaftNknjFBt5Zli19rDbBdV3s7+/Dtm30+30cHx/D\nMAy4rgvg8cpZURRteSsXIwH9pOuJ/L1cl4p/nkWasEdRtNKqWTLlS66PxW2S90jTdOo+1zRt7nO2\n+LNlVtGMsywLWZaVsppY8VgkSbLQNktibfx8sywLe3t7ODw8BPD4Gvjo0SP4vo8oimo3vomIiIio\n2raa/JEb83VO+5JVjvI8h+u66j3HAwJN09SUryzLVFPoIAgwGAzO/Xwcx+h0OnNvx6RgzrIslUSa\n9zWAzVQbFZu/yrL3juMgiiL0er1zgUmSJOh0OnN/lvEEkuu6ahwkSbJ0osYwDDQaDURRhCAI0Gq1\n4HkeGo0Gzs7O0O/3l3rddZuUUFvUtOBb9q1t2wiCAHmewzAMOI5T2aW29/f38fTTT2Nvbw9JkuDB\ngwc4Pj6G4zjI8xyDwQBxHJdahbFukkQBgF6vN/J380whleSd4zhI0xRnZ2e4du0aTNPEYDBAp9NZ\nOmGQ5znCMFTXOxGGoZoqaFnWygkJeR1JJFmWBdd1S0/OStJ+HaRH07wMw5j4877v46mnnsKtW7fg\neR46nQ5+/OMfYzgcYjAYIE3T2j9cICIiIqLq2HrljwQ9nudhOByW/h7SU0KqjIDJPTKkt0WapurJ\nrkxJmPaUd94AQKapjP/8ojf28jqbeCIsT+SLFU6yL6YFVYskEcaDdtnnqwY7pmmi1WohiqKRhJ5l\nWapapIoVQGV89lmvDUAleiS5WeXpU5IcCIJAfS8IAlU1IgmsKiauJpFpd6ZpTtzmeZJYlmXB9334\nvo8syxBFkZoal2UZBoPByLVhkWoc2abxbUuSBGEYwnXdmdUu4+8j59s4qULTdR1RFKn9YhhGqeel\nXMfXNT4WSf7ItXT880m1ZBzHqhedVHHJlGgmf4iIiIioLFuv/JGnmzI1p2yWZanKH5leJjfjxRtr\n+Z5UBkmvi2lNnyWpVLyhn1a94bouTNMcqRSS1wceB1gXJXR830ej0YBpmqqCZVogWYbx6RhSKVJc\nMa1IloIvHkOZOjceGEr1QrvdVkFOo9GA67pI0xTD4RC9Xk/t23mnwMjPXrt2DZ7nqSqX/f19ZFmm\nAtgHDx4ss0vWRj73uown7KSaa5Mrfrmuq86XeZKXMgbEYDBQ1RBSMSfVeVJZUaVEkJw/sk1yHZJE\nx8nJycjPS0XTrKl40pC+2Wyqc2Jvb08lgIIgGKlsmzf5Y9u2SkRMEgQBPM9Ds9lEGIYjlS+SEJZj\nIasguq577vVk+XbTNNXvy1j0PG/pZI1MWZOEDwA0Gg11/en1eiPXjzIqKBdJrMlUxfF/3+I4xmAw\nUONckneDwUBdb4tjXFZmXKU5+CLXUiIiIiLaLVuv/Ckmf5YlCZpJQeXh4SE8z1MJhiiKVCVI8Qba\nMAx4nqeCENM01RP2Xq+nppkAjwO5/f19pGmK09NTAFCJGbmhF81mE61WS1Up9Pt9NRVNbuyTJBkJ\nJoqNk6VfRLPZxN7ennqCHIah2mfFgEmSMFIxIY2a5YZ/vE+RJAbGAzUJNIHHwYtM+8qyTCWAivvb\n8zz11DpNUzXNyDTNkaBHAj3pIdTtdtFqtdBoNNQ0JNkHw+EQmqbBcRzEcYwgCC4MXCzLwnPPPaeq\nfx48eICDgwO1bweDAYIgQBiGI8Hipsmxl7FvWdZIVZemaWg0GgAeB42yKp4kFGTbi/tj2nkgx09e\nQ6bbyGsXf36Z6WdSJSdLYksy
}
],
"prompt_number": 76
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(template matching and peak finding)\n",
"\n",
"template matching:\n",
"\n",
"- template matching turns a difficult matching problem into an easier one\n",
"- the easier matching problem is peak finding\n",
"\n",
"peak properties:\n",
"\n",
"- peaks usually look roughly Gaussian (why?)\n",
"- absolute magnitude\n",
"- steepness / localization\n",
"- density / separation\n",
"\n",
"detect with:\n",
"\n",
"- Gaussian filtering and local maxima (scale space)\n",
"- mathematical morphology (peak templates)"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [],
"language": "python",
"metadata": {},
"outputs": []
}
],
"metadata": {}
}
]
}