hollance/Sketch.ipynb

## Sketch.ipynb
{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "https://github.com/keplerlab/neuralCreativityServer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "# The caffe module needs to be on the Python path; we'll add it here explicitly.\n",
    "import sys\n",
    "caffe_root = \"/Users/matthijs/Documents/Study/Machine_Learning/Tools/caffe/\"\n",
    "sys.path.insert(0, caffe_root + \"python\")\n",
    "import caffe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_def = 'deploy.prototxt'\n",
    "model_weights = 'sketchANet_full_iter_7500.caffemodel'\n",
    "\n",
    "if not os.path.isfile(model_weights):\n",
    "    print(\"Model not found\")\n",
    "    \n",
    "caffe.set_mode_cpu()\n",
    "net = caffe.Net(model_def, model_weights, caffe.TEST)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "labels_file = 'map.txt'\n",
    "if not os.path.exists(labels_file):\n",
    "    print(\"Labels not found\")\n",
    "    \n",
    "labels = np.loadtxt(labels_file, str, delimiter='\\t')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['0 airplane', '1 alarm_clock', '2 angel', '3 ant', '4 apple',\n",
       "       '5 arm', '6 armchair', '7 ashtray', '8 axe', '9 backpack'],\n",
       "      dtype='<U19')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x114a727f0>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQUAAAD8CAYAAAB+fLH0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGClMKBiGkcKEgmEYKUwoGIaR4v8D/hSjFkKgY/8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112915b70>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "img_path = 'cupresized/5121.png'\n",
    "img_path = 'cellphoneresized/3921.png'\n",
    "image = caffe.io.load_image(img_path)\n",
    "\n",
    "# The way the image gets loaded is as (128, 128, 3)\n",
    "# Here we convert it to grayscale\n",
    "# The background pixels are 1, the sketches are 0, so switch those around\n",
    "# Then multiply by 255 because the model expects values between 0 and 255\n",
    "image = 255*(1 - image[:,:,0])\n",
    "\n",
    "plt.imshow(image)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(128, 128)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "image.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predicted class is: 52\n",
      "Output label: 52 cell_phone\n"
     ]
    }
   ],
   "source": [
    "transformed_image = image\n",
    "\n",
    "net.blobs['data'].reshape(1,         # batch size\n",
    "                          1,         # 3-channel (BGR) images\n",
    "                          128, 128)  # image size is 227x227\n",
    "\n",
    "# Copy the image data into the memory allocated for the net\n",
    "net.blobs['data'].data[...] = transformed_image\n",
    "\n",
    "# Perform classification\n",
    "output = net.forward()\n",
    "\n",
    "# The output probability vector for the first image in the batch\n",
    "output_prob = output['prob'][0]\n",
    "\n",
    "print(\"Predicted class is:\", output_prob.argmax())\n",
    "print(\"Output label:\", labels[output_prob.argmax()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABBUAAAD8CAYAAAAohZxeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAFEBJREFUeJzt3X+Mndl5F/DvUxsXKQ2lsENU2d7YKq7A0CoJxokUVAJN\nwJtIdqv+wCsBCQo1SHEJJEI4UC3R8k9SRCohuVUXukqoSNwl0DIoBrdKgypQEzxpl6T2yu3I3dY2\noeumaQqqmq3Thz/mbriZndm5Pr723LE/H2nke8777L3Pro9ezX513nOruwMAAABwu75muxsAAAAA\ndiahAgAAADBEqAAAAAAMESoAAAAAQ4QKAAAAwBChAgAAADBEqAAAAAAMESoAAAAAQ4QKAAAAwJDd\n2/XBDz30UB84cGC7Ph4AAADYwKc//enf7O6lWWq3LVQ4cOBAVlZWtuvjAQAAgA1U1a/NWuvxBwAA\nAGCIUAEAAAAYIlQAAAAAhggVAAAAgCFCBQAAAGCIUAEAAAAYIlQAAAAAhggVAAAAgCFCBQAAAGDI\n7u1uAFgcB858bKa6Z9/3lrvcCQAAsBPYqQAAAAAMESoAAAAAQ4QKAAAAwBChAgAAADBEqAAAAAAM\nESoAAAAAQ4QKAAAAwJCZQoWqOlZVV6pqtarObHD94ar6RFX9YlV9pqrePP9WAQAAgEWyZahQVbuS\nnE3ySJLDSR6tqsPryn4gyVPd/eokJ5P88LwbBQAAABbLLDsVjiZZ7e6r3f18knNJTqyr6SR/ZPL6\n65P8r/m1CAAAACyi3TPU7E1ybWp8Pclr19W8N8lPV9X3J3lZkjfOpTsAAABgYc3roMZHk3ywu/cl\neXOSH6+qF713VZ2qqpWqWrl58+acPhoAAADYDrOECjeS7J8a75vMTXt7kqeSpLt/PskfTvLQ+jfq\n7ie6+0h3H1laWhrrGAAAAFgIs4QKF5McqqqDVbUnawcxLq+r+fUk354kVfWnsxYq2IoAAAAA97Et\nQ4XuvpXkdJILSZ7J2rc8XKqqx6vq+KTs3Um+r6r+Z5KPJHlbd/fdahoAAADYfrMc1JjuPp/k/Lq5\nx6ZeX07y+vm2BgAAACyyeR3UCAAAADxghAoAAADAEKECAAAAMESoAAAAAAwRKgAAAABDhAoAAADA\nEKECAAAAMESoAAAAAAwRKgAAAABDhAoAAADAEKECAAAAMESoAAAAAAwRKgAAAABDhAoAAADAEKEC\nAAAAMESoAAAAAAyZKVSoqmNVdaWqVqvqzAbXf6iqnp78/HJV/fb8WwUAAAAWye6tCqpqV5KzSd6U\n5HqSi1W13N2XX6jp7n8wVf/9SV59F3oFAAAAFsgsOxWOJlnt7qvd/XySc0lOvET9o0k+Mo/mAAAA\ngMU1S6iwN8m1qfH1ydyLVNUrkxxM8rObXD9VVStVtXLz5s3b7RUAAABYIPM+qPFkko9295c3utjd\nT3T3ke4+srS0NOePBgAAAO6lWUKFG0n2T433TeY2cjIefQAAAIAHwiyhwsUkh6rqYFXtyVpwsLy+\nqKr+VJJvSPLz820RAAAAWERbhgrdfSvJ6SQXkjyT5KnuvlRVj1fV8anSk0nOdXffnVYBAACARbLl\nV0omSXefT3J+3dxj68bvnV9bAAAAwKKb90GNAAAAwANCqAAAAAAMESoAAAAAQ4QKAAAAwBChAgAA\nADBEqAAAAAAMESoAAAAAQ4QKAAAAwBChAgAAADBEqAAAAAAMESoAAAAAQ4QKAAAAwBChAgAAADBE\nqAAAAAAMESoAAAAAQ2YKFarqWFVdqarVqjqzSc33VtXlqrpUVR+eb5sAAADAotm9VUFV7UpyNsmb\nklxPcrGqlrv78lTNoSTvSfL67v5CVf2Ju9UwAAAAsBhm2alwNMlqd1/t7ueTnEtyYl3N9yU5291f\nSJLufm6+bQIAAACLZpZQYW+Sa1Pj65O5ad+c5Jur6r9X1Ser6ti8GgQAAAAW05aPP9zG+xxK8oYk\n+5L8XFV9S3f/9nRRVZ1KcipJHn744Tl9NAAAALAdZtmpcCPJ/qnxvsnctOtJlrv797v7V5P8ctZC\nhq/S3U9095HuPrK0tDTaMwAAALAAZgkVLiY5VFUHq2pPkpNJltfV/FTWdimkqh7K2uMQV+fYJwAA\nALBgtgwVuvtWktNJLiR5JslT3X2pqh6vquOTsgtJPl9Vl5N8Isk/7O7P362mAQAAgO0305kK3X0+\nyfl1c49Nve4k75r8AAAAAA+AWR5/AAAAAHgRoQIAAAAwRKgAAAAADBEqAAAAAEOECgAAAMAQoQIA\nAAAwRKgAAAAADBEqAAAAAEOECgAAAMAQoQIAAAAwRKgAAAAADBEqAAAAAEOECgAAAMAQoQIAAAAw\nRKgAAAAADJkpVKiqY1V1papWq+rMBtffVlU3q+rpyc/fnn+rAAAAwCLZvVVBVe1KcjbJm5JcT3Kx\nqpa7+/K60p/o7tN3oUcAAABgAc2yU+FoktXuvtrdzyc5l+TE3W0LAAAAWHSzhAp7k1ybGl+fzK33\nXVX1mar6aFXtn0t3AAAAwMKa10GN/ynJge7+1iQ/k+RDGxVV1amqWqmqlZs3b87powEAAIDtMEuo\ncCPJ9M6DfZO5r+juz3f3lybDf53kz230Rt39RHcf6e4jS0tLI/0CAAAAC2KWUOFikkNVdbCq9iQ5\nmWR5uqCqvnFqeDzJM/NrEQAAAFhEW377Q3ffqqrTSS4k2ZXkye6+VFWPJ1np7uUkf6+qjie5leS3\nkrztLvYMAAAALIAtQ4Uk6e7zSc6vm3ts6vV7krxnvq0BAAAAi2xeBzUCAAAADxihAgAAADBEqAAA\nAAAMESoAAAAAQ4QKAAAAwBChAgAAADBEqAAAAAAMESoAAAAAQ4QKAAAAwBChAgAAADBEqAAAAAAM\nESoAAAAAQ4QKAAAAwBChAgAAADBEqAAAAAAMESoAAAAAQ2YKFarqWFVdqarVqjrzEnXfVVVdVUfm\n1yIAAACwiLYMFapqV5KzSR5JcjjJo1V1eIO6lyd5Z5JPzbtJAAAAYPHMslPhaJLV7r7a3c8nOZfk\nxAZ1/yzJ+5P83hz7AwAAABbULKHC3iTXpsbXJ3NfUVWvSbK/uz/2Um9UVaeqaqWqVm7evHnbzQIA\nAACL444Paqyqr0nygSTv3qq2u5/o7iPdfWRpaelOPxoAAADYRrOECjeS7J8a75vMveDlSf5skv9a\nVc8meV2SZYc1AgAAwP1tllDhYpJDVXWwqvYkOZlk+YWL3f3F7n6ouw9094Ekn0xyvLtX7krHAAAA\nwELYMlTo7ltJTie5kOSZJE9196Wqeryqjt/tBgEAAIDFtHuWou4+n+T8urnHNql9w523BQAAACy6\nOz6oEQAAAHgwCRUAAACAIUIFAAAAYIhQAQAAABgiVAAAAACGCBUAAACAIUIFAAAAYIhQAQAAABgi\nVAAAAACGCBUAAACAIUIFAAAAYIhQAQAAABgiVAAAAACGCBUAAACAIUIFAAAAYMhMoUJVHauqK1W1\nWlVnNrj+d6vqs1X1dFX9t6o6PP9WAQAAgEWyZahQVbuSnE3ySJLDSR7dIDT4cHd/S3e/KskPJvnA\n3DsFAAAAFsosOxWOJlnt7qvd/XySc0lOTBd09+9MDV+WpOfXIgAAALCIds9QszfJtanx9SSvXV9U\nVe9I8q4ke5L85bl0BwAAACysuR3U2N1nu/ubkvyjJD+wUU1VnaqqlapauXnz5rw+GgAAANgGs4QK\nN5Lsnxrvm8xt5lyS79joQnc/0d1HuvvI0tLS7F0CAAAAC2eWUOFikkNVdbCq9iQ5mWR5uqCqDk0N\n35LkV+bXIgAAALCItjxTobtvVdXpJBeS7EryZHdfqqrHk6x093KS01X1xiS/n+QLSd56N5sGAAAA\ntt8sBzWmu88nOb9u7rGp1++cc18AAADAgpvbQY0AAADAg0WoAAAAAAwRKgAAAABDhAoAAADAEKEC\nAAAAMESoAAAAAAwRKgAAAABDhAoAAADAEKECAAAAMESoAAAAAAwRKgAAAABDhAoAAADAEKECAAAA\nMESoAAAAAAwRKgAAAABDZgoVqupYVV2pqtWqOrPB9XdV1eWq+kxVfbyqXjn/VgEAAIBFsmWoUFW7\nkpxN8kiSw0kerarD68p+McmR7v7WJB9N8oPzbhQAAABYLLPsVDiaZLW7r3b380nOJTkxXdDdn+ju\n350MP5lk33zbBAAAABbNLKHC3iTXpsbXJ3ObeXuS/3wnTQEAAACLb/c836yq/nqSI0n+4ibXTyU5\nlSQPP/zwPD8aAAAAuMdm2alwI8n+qfG+ydxXqao3JvknSY5395c2eqPufqK7j3T3kaWlpZF+AQAA\ngAUxS6hwMcmhqjpYVXuSnEyyPF1QVa9O8qNZCxSem3+bAAAAwKLZMlTo7ltJTie5kOSZJE9196Wq\neryqjk/K/nmSr0vy76rq6apa3uTtAAAAgPvETGcqdPf5JOfXzT029fqNc+4LAAAAWHCzPP4AAAAA\n8CJCBQAAAGCIUAEAAAAYIlQAAAAAhggVAAAAgCFCBQAAAGCIUAEAAAAYIlQAAAAAhggVAAAAgCFC\nBQAAAGCIUAEAAAAYIlQAAAAAhggVAAAAgCFCBQAAAGCIUAEAAAAYIlQAAAAAhswUKlTVsaq6UlWr\nVXVmg+vfVlW/UFW3quq7598mAAAAsGi2DBWqaleSs0keSXI4yaNVdXhd2a8neVuSD8+7QQAAAGAx\n7Z6h5miS1e6+miRVdS7JiSSXXyjo7mcn1/7gLvQIAAAALKBZHn/Ym+Ta1Pj6ZO62VdWpqlqpqpWb\nN2+OvAUAAACwIO7pQY3d/UR3H+nuI0tLS/fyowEAAIA5myVUuJFk/9R432QOAAAAeIDNEipcTHKo\nqg5W1Z4kJ5Ms3922AAAAgEW3ZajQ3beSnE5yIckzSZ7q7ktV9XhVHU+SqvrzVXU9yfck+dGqunQ3\nmwYAAAC23yzf/pDuPp/k/Lq5x6ZeX8zaYxEAAADAA+KeHtQIAAAA3D+ECgAAAMCQmR5/AADm58CZ\nj81c++z73nIXOwEAuDN2KgAAAABDhAoAAADAEKECAAAAMESoAAAAAAwRKgAAAABDhAoAAADAEF8p\nCTywfK0fAADcGTsVAAAAgCFCBQAAAGCIUAEAAAAYIlQAAAAAhjioEQAA4C5yODT3s5l2KlTVsaq6\nUlWrVXVmg+tfW1U/Mbn+qao6MO9GAQAAgMWy5U6FqtqV5GySNyW5nuRiVS139+Wpsrcn+UJ3/8mq\nOpnk/Un+2t1oGAAAgPufHR47wyw7FY4mWe3uq939fJJzSU6sqzmR5EOT1x9N8u1VVfNrEwAAAFg0\ns5ypsDfJtanx9SSv3aymu29V1ReT/PEkvzmPJmGnmTVVlajuHNN/p/7eAABgzT09qLGqTiU5NRn+\n36q6ci8//y56KAIUBtT7t7uDr7itNbxAfW+LB/3ff0Hdt/dh6+2Bcd+uYR4Y1vCcuO+PmcN/N2v4\nq71y1sJZQoUbSfZPjfdN5jaquV5Vu5N8fZLPr3+j7n4iyROzNrdTVNVKdx/Z7j5glDXMTmcNs9NZ\nw+x01jA7nTU8bpYzFS4mOVRVB6tqT5KTSZbX1Swneevk9Xcn+dnu7vm1CQAAACyaLXcqTM5IOJ3k\nQpJdSZ7s7ktV9XiSle5eTvJjSX68qlaT/FbWggcAAADgPjbTmQrdfT7J+XVzj029/r0k3zPf1naU\n++6RDh441jA7nTXMTmcNs9NZw+x01vCg8pQCAAAAMGKWMxUAAAAAXkSocIeq6lhVXamq1ao6s939\nwFaq6tmq+mxVPV1VK5O5P1ZVP1NVvzL58xu2u0+YVlVPVtVzVfVLU3Mbrtta8y8n9+XPVNVrtq9z\nWLPJGn5vVd2Y3I+frqo3T117z2QNX6mqv7o9XcOaqtpfVZ+oqstVdamq3jmZdx9mR3iJNew+PAdC\nhTtQVbuSnE3ySJLDSR6tqsPb2xXM5C9196umvjbnTJKPd/ehJB+fjGGRfDDJsXVzm63bR5Icmvyc\nSvIj96hHeCkfzIvXcJL80OR+/KrJGVaZ/C5xMsmfmfwzPzz5nQO2y60k7+7uw0lel+Qdk3XqPsxO\nsdkaTtyH75hQ4c4cTbLa3Ve7+/kk55Kc2OaeYMSJJB+avP5Qku/Yxl7gRbr757L27ULTNlu3J5L8\nm17zySR/tKq+8d50ChvbZA1v5kSSc939pe7+1SSrWfudA7ZFd3+uu39h8vr/JHkmyd64D7NDvMQa\n3oz78G0QKtyZvUmuTY2v56UXJyyCTvLTVfXpqjo1mXtFd39u8vp/J3nF9rQGt2WzdevezE5yerI9\n/MmpR8+sYRZWVR1I8uokn4r7MDvQujWcuA/fMaECPHj+Qne/JmtbE99RVd82fbHXvhLG18Kwo1i3\n7FA/kuSbkrwqyeeS/IvtbQdeWlV9XZJ/n+Tvd/fvTF9zH2Yn2GANuw/PgVDhztxIsn9qvG8yBwur\nu29M/nwuyU9mbSvXb7ywLXHy53Pb1yHMbLN1697MjtDdv9HdX+7uP0jyr/L/t9ZawyycqvpDWfuf\nsX/b3f9hMu0+zI6x0Rp2H54PocKduZjkUFUdrKo9WTvMY3mbe4JNVdXLqurlL7xO8leS/FLW1u1b\nJ2VvTfIft6dDuC2brdvlJH9zcvr465J8cWp7LiyMdc+Yf2fW7sfJ2ho+WVVfW1UHs3bY3f+41/3B\nC6qqkvxYkme6+wNTl9yH2RE2W8Puw/Oxe7sb2Mm6+1ZVnU5yIcmuJE9296VtbgteyiuS/OTafTW7\nk3y4u/9LVV1M8lRVvT3JryX53m3sEV6kqj6S5A1JHqqq60n+aZL3ZeN1ez7Jm7N2qNLvJvlb97xh\nWGeTNfyGqnpV1raMP5vk7yRJd1+qqqeSXM7aieXv6O4vb0ffMPH6JH8jyWer6unJ3D+O+zA7x2Zr\n+FH34TtXa48/AQAAANwejz8AAAAAQ4QKAAAAwBChAgAAADBEqAAAAAAMESoAAAAAQ4QKAAAAwBCh\nAgAAADBEqAAAAAAM+X8XMhh40MZ2KQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112915710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(18,4))\n",
    "plt.bar(range(0, 256), output_prob, width=3)\n",
    "plt.xticks(range(0, 256, 50))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "52 0.851141 52 cell_phone\n",
      "114 0.067519 114 ipod\n",
      "69 0.038553 69 door\n",
      "222 0.018700 222 telephone\n",
      "244 0.012915 244 walkie_talkie\n"
     ]
    }
   ],
   "source": [
    "top_ixs = output_prob.argsort()[::-1][:5]\n",
    "for i in top_ixs:\n",
    "    print(\"%d %f %s\" % (i, output_prob[i], labels[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}