nvidia-AI.summerschool17-no.../teaching-lw/020-frequency-domain.ipynb

772 lines
1.9 MiB
Text
Raw Permalink Normal View History

2017-09-22 17:26:46 +00:00
{
"metadata": {
"name": "020-frequency-domain"
},
"nbformat": 3,
"nbformat_minor": 0,
"worksheets": [
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Title: Introduction to the Frequency Domain\n",
"Author: Thomas Breuel\n",
"Institution: UniKL"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Introduction to the Frequency Domain\n",
"===================================="
]
},
{
"cell_type": "code",
"collapsed": true,
"input": [
"\n",
"from scipy.ndimage import filters\n",
"def imrow(*args,**kw):\n",
" gray()\n",
" size = kw.get(\"s\",6)\n",
" if \"s\" in kw: del kw[\"s\"]\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,interpolation='nearest',**kw)\n",
"def roll2(image,dx,dy):\n",
" return roll(roll(image,dx,axis=0),dy,axis=1)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 1
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Periodic Boundary Conditions\n",
"----------------------------\n",
"\n",
"Fourier transforms use periodic boundary conditions (we had talked about this before).\n",
"So, in order to make our signal looks nice and periodic,\n",
"we modify the image so that it goes smoothly towards zero around the edges of the image.\n",
"In practice, we don't need to do something as drastic as this,\n",
"but in the example, the intermediate results look nicer."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"image = mean(imread(\"bridge.jpg\"),axis=2)\n",
"image = image-mean(image)\n",
"w,h = image.shape\n",
"ys,xs = meshgrid(range(-h//2,h-h//2),range(-w//2,w-w//2))\n",
"r = minimum(h//2-abs(ys),w//2-abs(xs))\n",
"r = 1.0-exp(-r**2/80.0**2)\n",
"image = r*image\n",
"imrow(image)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAXMAAAEZCAYAAABl1cWuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nXjetwNQBns+Z+irvyd8AlJIx
}
],
"prompt_number": 2
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"plot(image[200,:])"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 3,
"text": [
"[<matplotlib.lines.Line2D at 0x36b3d90>]"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAD9CAYAAABeOxsXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n3ZjBU08Bjz8eejs3EcsgcTFb
}
],
"prompt_number": 3
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Fourier Transforms and the Frequency Domain\n",
"------------------"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The Fourier transform is a linear, invertible transformation of the image\n",
"that yields a complex result.\n",
"The Fourier transform for images is implemented by the `fft2` function,\n",
"and its inverse by the `ifft2` function.\n",
"\n",
"In principle, `ifft2(fft2(image))` should be the identity function,\n",
"but `ifft2` returns complex numbers; the imaginary part is nearly zero (roundoff error),\n",
"so we are just interested in the real part.\n",
"\n",
"Note that the Fourier transformation is _not shift invariant_."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"f = fft2(image)\n",
"imrow(image,real(ifft2(f)))\n",
"print amax(abs(imag(ifft2(f))))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "stream",
"stream": "stdout",
"text": [
"6.03959719156e-13\n"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAECCAYAAAAfJQrzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nllU76sfr9dadbVkqlerKz+uH
}
],
"prompt_number": 4
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"What is this mysterious Fourier transform `f`?\n",
"\n",
"We can't really easily tell, because it is complex.\n",
"However, the magnitude is kind of instructive.\n",
"Since the magnitude spans a very large range,\n",
"we compress the range with a $\\gamma=0.2$."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(image,abs(f),abs(f)**0.2)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABBAAAAD7CAYAAADTqWeiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n1TAejxEKhaSNgsGg9D8CmPP5
}
],
"prompt_number": 5
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's try some simple test images. The most obvious one is the impulse."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"impulse = zeros((256,256))\n",
"impulse[128,128] = 1.0\n",
"spectrum = abs(fft2(impulse))**0.2\n",
"imrow(impulse,spectrum)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsMAAAFOCAYAAABjShDZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmZkppZRy+fLl0tfXV06fPl1e
}
],
"prompt_number": 6
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Didn't that work? Actually, it did, but the output is constant."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"plot(spectrum[10,:])"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 7,
"text": [
"[<matplotlib.lines.Line2D at 0x4f94990>]"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAX8AAAD9CAYAAABUS3cAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\neI14z7+8vFy9vb0Kh8OKxWLa
}
],
"prompt_number": 7
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"It's only the magnitude of the output that's constant. The real part jumps back between -1 and 1."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"f = fft2(impulse)[50,50:100]\n",
"plot(f.real)\n",
"plot(f.imag)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "pyout",
"prompt_number": 8,
"text": [
"[<matplotlib.lines.Line2D at 0x2b95f90>]"
]
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD9CAYAAABQvqc9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nF77wBezZsyfx+1T+Z+U2p9UP
}
],
"prompt_number": 8
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Maybe something slightly bigger than an impulse gives us a better idea."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"step = zeros((256,256))\n",
"step[112:144,112:144] = 1.0\n",
"spectrum = abs(fft2(step))**0.2\n",
"imrow(step,spectrum)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsMAAAFOCAYAAABjShDZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxYMPPog77rijsgGmlyM89NBD
}
],
"prompt_number": 9
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"That's still pretty mysterious. \n",
"\n",
"Actually, it turns out that the signal with the simplest 2D Fourier transform is a sine wave.\n",
"This will become clearer when we look at this again in processing of sound signals later."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"ys,xs = meshgrid(range(300),range(300))\n",
"wave = sin((xs+2.0*ys)*pi/15.0)\n",
"spectrum = abs(fft2(wave))**0.2\n",
"imrow(wave,spectrum,spectrum[:50,:50])"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABBAAAAFFCAYAAABVHh4TAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n2BMnTujEiRN116nWcc3MGfVY
}
],
"prompt_number": 10
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Notice that there is just a single pixel at (10,20). Why is it there? Roughly speaking, because there are 30 repeats in the y direction, and 60 repeats in the x direction.\n",
"\n",
"We can now mix multiple of these signals and the Fourier transform will resolve these components for us."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"wave = sin((xs+2.0*ys)*pi/15.0)+cos((3.0*xs+ys)*pi/15.0)+sin((xs+ys)*pi/15.0)+3.0*sin(xs*pi/150.0)+3.0*sin(ys*pi/150.0)\n",
"spectrum = abs(fft2(wave))**0.2\n",
"imrow(wave,spectrum,spectrum[:40,:40])"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABBAAAAFFCAYAAABVHh4TAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n2uW+5N3xfeU9kXfH+5U9KW56
}
],
"prompt_number": 11
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Fourier transforms model images as being composed of sine waves that are shifted and added.\n",
"\n",
"This turns out not to be a very good statistical model of real-world images.\n",
"It is, however, a good model for audio and radio signals, which is where most\n",
"of the theory comes from. And the theory works even if the images are\n",
"not naturally represented using the basis functions of the filters."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Frequency Domain Filtering\n",
"-------------------------"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The basic idea behind frequency domain filtering is to perform a Fourier transform,\n",
"select a subset of the frequencies--those we are interested in--and then transform back.\n",
"\n",
"Now, the low frequencies happen to be in the corners of the Fourier transform (this\n",
"is just a convention). \n",
"\n",
"So, we take the original Fourier transform and just copy values from its corners,\n",
"then transform back. And let's also transform back everything but the values we selected."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"f = fft2(image)\n",
"f2 = zeros(f.shape,dtype=complex)\n",
"d = 30\n",
"f2[:d,:d] = f[:d,:d]\n",
"f2[:d,-d:] = f[:d,-d:]\n",
"f2[-d:,:d] = f[-d:,:d]\n",
"f2[-d:,-d:] = f[-d:,-d:]\n",
"imrow(abs(f)**0.2,abs(f2)**0.2,abs(f-f2)**0.2)\n",
"imrow(image,real(ifft2(f2)),real(ifft2(f-f2)))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABBAAAAD7CAYAAADTqWeiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nHdPpVPJo7utmsxmr1QqqqiIc
},
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAABBAAAAD7CAYAAADTqWeiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nu96728ZcDO3s9XrY2toKH16z
}
],
"prompt_number": 12
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"What we can see here is that if we just select low frequencies (a *low pass filter*),\n",
"we get a filter that looks like a smoothing filter.\n",
"\n",
"If we just select high frequencies (a *high pass filter*,\n",
"we get a filter that looks like a derivative filter.\n",
"\n",
"However, there is a lot of *ringing* in these images,\n",
"the little wavy black lines surrounding edges in the filtered images.\n",
"Strict low-pass or high-pass filters show this effect."
]
},
{
"cell_type": "heading",
"level": 1,
"metadata": {},
"source": [
"Filters in the Frequency Domain"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(Convolution vs Fourier Transform)\n",
"\n",
"The Fourier transform is a very special kind of filter,\n",
"because it transforms convolutions into multiplications.\n",
"\n",
"Let's write ${\\cal F}[s]$ for the Fourier transform of\n",
"the signal $s$. Then,\n",
"\n",
"${\\cal F}[s * t] = {\\cal F}[s] \\cdot {\\cal F}[t]$\n",
"\n",
"or\n",
"\n",
"$s * t = {\\cal F}^{-1}[{\\cal F}[s] \\cdot {\\cal F}[t]]$\n",
"\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"(Speedup from FFT)\n",
"\n",
"This matters a great deal because using the *fast Fourier transform* algorithm (*FFT*),\n",
"we can compute the Fourier transform of a signal with $N$ elements in time $N \\log N$.\n",
"\n",
"So, a naive implementation of convolution of two sequences of length $N$ takes time $O(N^2)$,\n",
"while a convolution implemented with Fourier transforms takes time $O(N \\log N)$.\n",
"Radios, disk drive, cell phones, and many other modern devices take advantage of this."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let us explore this issue computationally a bit more by looking at the Gaussian filter\n",
"in the image domain and the frequency domain.\n",
"\n",
"\n"
]
},
{
"cell_type": "code",
"collapsed": true,
"input": [
"def impulse(w,h):\n",
" image = zeros((w,h))\n",
" image[w//2,h//2] = 1.0\n",
" return image"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 13
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Let's compute the impulse response of the Gaussian filter."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"from scipy.ndimage import filters\n",
"gaussian = filters.gaussian_filter(impulse(64,64),8.0)\n",
"imrow(gaussian)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAAFqCAYAAAA+xmCkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0Pc9jvse+QaMrp+PuJ+lQoWO
}
],
"prompt_number": 14
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can convolve with that directly to obtain the following image."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"result1 = filters.convolve(image,gaussian)\n",
"imrow(result1)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAXMAAAEZCAYAAABl1cWuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhWq0n3Vs0F/O69rd3Y2I90Mr
}
],
"prompt_number": 15
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Alternatively, we can multiply with the Fourier transform of the impulse response.\n",
"\n",
"(We recompute the Fourier transform because the `fft2` function wants the impulse\n",
"response centered on (0,0), while the `convolve` function wants it centerd on `(w//2,h//2)`.\n",
"Furthermore, `fft2` requires an impulse response padded with zeros to the size of the image.)"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"gaussian0 = filters.gaussian_filter(impulse(*image.shape),8.0)\n",
"gaussian = roll2(gaussian0.copy(),w//2,h//2)\n",
"imrow(gaussian0,gaussian)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAECCAYAAAAfJQrzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nS7ZTk7hADgAAmxCCERjLrQmu
}
],
"prompt_number": 16
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"result2 = real(ifft2(fft2(image)*fft2(gaussian)))\n",
"imrow(result2)"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAXMAAAEZCAYAAABl1cWuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n2AG5gnBmpPjTCtVoP+vYoL+c
}
],
"prompt_number": 17
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"These two are the same.\n",
"\n",
"Since we're going to look at more spectra (magnitudes of Fourier transforms), let's\n",
"define a function that makes this more convenient.\n",
"This function also shifts the low frequencies to the center of the image, where\n",
"they are easier to see."
]
},
{
"cell_type": "code",
"collapsed": true,
"input": [
"def spectrum(image,mode=lambda x:x**0.2,eps=1e-6):\n",
" w,h = image.shape\n",
" return roll2(mode(maximum(eps,abs(fft2(image)))),w//2,h//2)"
],
"language": "python",
"metadata": {},
"outputs": [],
"prompt_number": 18
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(image,spectrum(image))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAECCAYAAAAfJQrzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZDKxFHkoFDLWFsC2XC7VarVU
}
],
"prompt_number": 19
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"So, the spectrum of the original image has a lot of high frequency components.\n",
"These are organized in \"spikes\", and these are actually perpendicular to the\n",
"largest edges found in the image.\n",
"\n",
"Let's look at the filtered outputs."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(result1,spectrum(result1))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAECCAYAAAAfJQrzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nsWEOVriCBb7sBHvj8FruJDrL
}
],
"prompt_number": 20
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The above filter was implemented by the `convolve` function.\n",
"It got rid of most of the high frequency components, although there are\n",
"smallish streaks still visible."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(result2,spectrum(result2))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAECCAYAAAAfJQrzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nuCKc7CG16XQaETF3fhfL8kmB
}
],
"prompt_number": 21
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"That filter was implemented by Fourier transform. The image looks the same, but it really\n",
"got rid of all high frequency components. This difference is actually very slight and\n",
"probably related to the tructation in the `convolve` function. It is only visible because\n",
"we are using $\\gamma=0.2$ for display.\n",
"\n",
"Now, we multiply the image with the Fourier transform of the Gaussian.\n",
"What does that look like?"
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(roll2(gaussian,w//2,h//2),spectrum(gaussian))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAECCAYAAAAfJQrzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nbtXBgwc1NTVVX+fEiRPasmWL
}
],
"prompt_number": 22
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The smooth falloff of the Gaussian is really important. \n",
"Let's construct a lowpass filter of exactly the same size as the Gaussian and compare the two."
]
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"f = fft2(gaussian)\n",
"mag = abs(f)\n",
"lowpass = where(mag>0.5,f/mag,0.0)\n",
"imrow(roll2(abs(f),w//2,h//2),roll2(abs(lowpass),w//2,h//2))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAECCAYAAAAfJQrzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nODuE868XnL03AMDLe5n3VnBq
}
],
"prompt_number": 23
},
{
"cell_type": "code",
"collapsed": false,
"input": [
"imrow(real(ifft2(fft2(image)*f)),real(ifft2(fft2(image)*lowpass)))"
],
"language": "python",
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"png": "iVBORw0KGgoAAAANSUhEUgAAAsEAAAECCAYAAAAfJQrzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzsfU2opVl59b5/VV1th89Rd1tVDT0wJPYgP7TaDhJwoAYn\nIghtFKIkQmaR4EClh5nYjkIkOAhoaAiIzhQCGThxENBWEkKgBzpR2mrtEMR01/+te+83aNa566yz\n1rP3e+65Vfd2vQ8cznven/377PWs59n73Wfr6OjoqM0yyyyzzDLLLLPMMstDJNsPugCzzDLLLLPM\nMssss8xyv2UmwbPMMssss8wyyyyzPHQyk+BZZplllllmmWWWWR46mUnwLLPMMssss8wyyywPncwk\neJZZZplllllmmWWWh05mEjzLLLPMMssss8wyy0Mnp0KC/+3f/q39/u//fvvd3/3d9tWvfvU0sphl\nlllmmWVDMmP2LLPM8jDK1qb3CT44OGi/93u/177//e+3K1eutPe9733tW9/6VnvPe96zyWxmmWWW\nWWbZgMyYPcssszyssvFI8Msvv9ze/e53t6effrrt7e21P//zP2/f/e53N53NLLPMMsssG5AZs2eZ\nZZaHVTZOgq9du9aeeuqpxe+rV6+2a9eubTqbWWaZZZZZNiAzZs8yyywPq+xuOsGtra2N3DPLLLPM\ncpbl7fKP8zNmzzLLLA+DOMzeOAm+cuVKe/XVVxe/X3311Xb16tWV+z7/+c9vOusVOS3g/uEPf9g+\n8IEPnErapy3nueytzeV/kHKaZdexehoE84c//GF77rnnNpLW1772tY2kcxZkFLOfffbZ9uyzz7bW\ncv+Mntf+xm/+xgeyvb29cl8lyPPo6Ki9/PLL7b3vfW+8XpVdhfNOx71y4dNaa4eHh+3o6KgdHh6u\n3IPj//qv/2p/+Id/OFSWde3e9vbxxHCVbq/Oeu7HP/5xe9/73hfzcs9pX3Db8PWT9N3ovT/5yU8W\nuuOed3VhcWVP5Xf1cOd67YFj1pvUjzqueNylcVqJG6c6llMdeWwcHh62H//4x4u253bTuibhPP/p\nn/7J3rNxEvze9763/exnP2s///nP2+XLl9u3v/3t9q1vfWtyOnPkYZZKzqt+9MDgLMv9LHsvn6kk\n+e0StT0NWQez0T+J3PZIb3Vtqo65+4+OjtrW1tbiu7W3jLOSB76/dzxS3lQWPmaDj/LoeffM4eHh\nSl5cv0p6pLOqx0lJ/9bWViTY1TOtHbed9t2Uuq8rqT+0jEmnUnpOqvo452BnZ2fJiXJBhER6lfy6\nD2RKv6luVM6T1oeP0Ra9+1xd15GNk+Dd3d32j//4j+3P/uzP2sHBQfvc5z43/JbxeSUHpyWn0R7n\nmYS1dv7Lf1pyVsjpSWQKUXVAeZJ0p0YV3k6yScxmQ+bOV8+fhfZdJzI3xWFz0V4XGUx5a55VpA0E\nZiSq13NETkK0XdlGooK4ryLCI7IJh5mdqiSjdUr1G2kj5xyBEDMZZlKdZlT4g77TayN1U5nynNY/\nORxcX20H/r0OhmycBLfW2kc/+tH20Y9+dNIzU7zDBy1PPfXUmSnLVHHTnA9Sprbj1atXz23bt3a+\ny3/aupPaZd0orj535cqVE6V52pGnBykjmH358uWuYV6nfZzRHCGWKershPue86pILkfaNiFpGnxk\nSv+JJ55YKnciv0x8e5FAPa6kN+Wf0tra2mpXrlzZ2LhJ0fxKL6fgijv3rne9K16bKr2ZhZG+STq0\nvb29QiKffPLJJZKrxJejwo4cb0pYf0ZmY3DP5cuXS0exktHynwoJ3pScVYMzSgbOYvn5LfDzKHP5\nH5w8qLJvKup75cqVExuytzMR7snly5ft+dEAxkg0mH8nYqPkd4TkOBJ8muIif71vFW4HkJleZE+f\nS23Zm+bmcyclHRVuVNPeSaplLSPlUb3p5fmud73LOnopolvln8rkzrs+6s0q6DG3vdOfnZ2dSIpH\nI9u9vqgIsM6GaD3Q9m6JkLZNJdU9Z4IEjyrLw2p8ZunLSKRilgcnm4qmQUaMZ2XcTrL8YRMRofMo\nJ8XfdadUIYeHh4toF99z0n5GWtWaxHXSroiw5p3OKWlRsuKup8iwS7sqg6tHqsNInXoyJZ+pL6Ox\nOCeq6vtN846UXhXFd2XD/SCISopTvhrxTRHhJLw0BBHo1nKbu/NK2PE5ODiw508SBe7135kgwSrr\nTNfMcnI5j0TyYdOP+1HfTURLVUZ066QvLVVrydaRk6wze7vLg2oT6AiT4anPqjgSxERpyhQ8nuHr\no7MHVQR3yktMLiLsvqs6cF1GI6ZTZdRJ4LavcITrwi+OpbR6Mw2pPCnPdG6E9OK7cmicMElEnVNb\ncno6e8C/neOk6alejGA3/3aRa+ySkghwa7UurKPfrZ0BElwpzP0A2vNI/E5TzrLBP8tlG5HzUv6p\nU6FV5GRThHrEYG4iIgjROpyXvjtNGW2DUUw9yezApmYWlPSm60qqRiKN+runR0poE+kdWfebzrGM\nRFNPazakGs9V22I2oLU+2XRR3TQ1z22VptpHpuCr/q2cFHdcvdjIbeYiwEoeU35Jz0ZEI9Daj1Uf\nu3In8ltFmEedoqpOD5wEs/QKnSr8djNQ560+97u8Z7l9znLZKklGrYoC9aal9fdU4pKiBpp2Kvs6\nRImnGM9rX95vqabcR56r+nLTy2ic9EiGK9vOzk6Z5hSS6CKArfVfVuqR2KlRSq0r7zqAe6tlA6OS\nxvKUvp4S3U645GYTNhHtTm1ckV+Ux/W1q5trQ3fs2jRFoCHAPtUH/ujWfj3yuk55XV/0cGWdaPCZ\nIsEQLfTIXnXn0WCdlzI/DCT3vPTF/ZYe0dX79HhdEqMAuQnDOSIcXdApv1lHVqU3Ba/HLI5Q8fHR\nUX+9Yc95S+XVe106VSR0k9Jb6uDWAHPZW+uvie/ZTafrozMrI/n00hkdx+4lwF7E2xE8V/Yp5U3p\np/OpjGl9d4rI6nhwpJjxMunHiKNQRW9xjPuwlteVwaVVEWQ+r22YcPgkvPBMkmCW0SjDWTJQZ6ks\nTmaSeTI5j0toekZmxDCMTHOlvKYYDwegI9NjvfKP5t+bZpvFS4+ouftHIosgwydxeFwgZcrMY6VT\nU7FN83WEd4QQOeFxk9o02c/euKlmY/Se08D7KlI6mt+DtEMja7ZdXRwH6v1ZRurPdWfSUsS3Wsdb\n7V7h8nZE3RHhJD1nKMmZIcFTK+c6/bTlLBK5s1imdeUkxONBtsNZ6oMEaNq2I4Qikd2RaGxlKPke\n13aJAPemykbIbyL7Z6kPz6qMRMNdZK6H7UqEkwF3RNjdW5GIioCsK1OfVUI+GgF2os6hi3q6CJum\nWUUC3bMncQpcn1dOjtvj1rXVSN5V2Uc4yNTIsUt7VAdHtrHTttNyclv3ggjJyXEY7EiwWyLhfldl\n4Loz3lROatVvI5h1ZkgwpBfWHvXiT5LXwySnGfHaRPuep769n3mOgvFIpE3vHfnu/Xf9qLF0IOUA\nNL180cuD83L38vlEyGecOBZtjxQ1n0IuRyLFELc8osqjN31eGdfR8iaCUp3T9klEL6XlCIYjkm6c\nuLGgx0qYHJHpSUXcR9vdRUnxGVkmonUbLe+UMX+SNqny0aVZ6R4ti+tLR1LdMb5Ho8DVdS3fiP6g\nPXgXGL3m7u8R4UrOHAlm0f3qRhT+QcuDmEY97fpvMv37RWrfLiS4ilq46M1IxGKEwI4CY3q+V4ZE\ngDntRH7XIcBTye+8JGJV0FaVY1Bdm7IbgYsWprRUn6utp9R2pGhrLxgzen8qc0XKU8CnN74SKdF7\n+bgixI40uXxdnfjb3ae6lPTGtU+1o8FIP4zqaC+tKZHNdWXKjF2Kvo4sWagcn96fcozc575Z0niY\nKus8c6ZJMKTy5FU2YbjuN4E6TyR2k2meJrG9n324ibx6hmU0osO/9Xvkbd30TO8+l0dP3JSXI929\nPKeKI1hnyZE+q5IcCbdtVdWm
}
],
"prompt_number": 24
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"As we can see above, the Gaussian-filtered image looks smoothly blurred,\n",
"while the lowpass filtered image has roughly the same amount of blur,\n",
"but lots of additional \"structure\" due to ringing."
]
}
],
"metadata": {}
}
]
}