weighting windows.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import torch\n",
    "import numpy as np\n",
    "from PIL import Image\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "from pytorch_sift import SIFTNet\n",
    "sift_desc = SIFTNet(65)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f424782ccd0>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVUAAAFSCAYAAABCCUyWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhSmC8qxVtgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f424abd1a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVUAAAFSCAYAAABCCUyWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f42b969ee90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "mp_gauss = sift_desc.CircularGaussKernel(41, circ = True, sigma_type = 'hesamp')\n",
    "plt.figure()\n",
    "plt.grid(False)\n",
    "plt.imshow(255-mp_gauss, interpolation='None')\n",
    "\n",
    "vlfeat_gauss = sift_desc.CircularGaussKernel(41, circ = False, sigma_type ='vlfeat')\n",
    "plt.figure()\n",
    "plt.grid(False)\n",
    "plt.imshow(255-vlfeat_gauss, interpolation='None')\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}