Instant Recognition with Caffe.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib as mpl\n",
    "%matplotlib inline \n",
    "\n",
    "plt.rcParams['figure.figsize'] = (10,10)\n",
    "plt.rcParams['image.interpolation'] = 'nearest'         \n",
    "plt.rcParams['image.cmap'] = 'gray'\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/Bato/Documents/caffe/python/caffe/pycaffe.py:13: RuntimeWarning: to-Python converter for boost::shared_ptr<caffe::Net<float> > already registered; second conversion method ignored.\n",
      "  from ._caffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, \\\n",
      "/Users/Bato/Documents/caffe/python/caffe/pycaffe.py:13: RuntimeWarning: to-Python converter for boost::shared_ptr<caffe::Blob<float> > already registered; second conversion method ignored.\n",
      "  from ._caffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, \\\n",
      "/Users/Bato/Documents/caffe/python/caffe/pycaffe.py:13: RuntimeWarning: to-Python converter for boost::shared_ptr<caffe::Solver<float> > already registered; second conversion method ignored.\n",
      "  from ._caffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, \\\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "caffe_root = '/Users/Bato/Documents/caffe/'\n",
    "sys.path.insert(0, caffe_root+'python')\n",
    "\n",
    "import caffe\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "caffe.set_mode_cpu()\n",
    "\n",
    "model_def = caffe_root + 'models/fine_tune_CaffeNet/Tune_fc78_1500iters/test2_4/deploy.prototxt'\n",
    "model_weight = caffe_root + 'models/fine_tune_CaffeNet/Tune_fc78_1500iters/test2_4/caffenet_train_iter_1500.caffemodel'\n",
    "\n",
    "net = caffe.Net(model_def, model_weight, caffe.TEST)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Set up input preprocessing. (We'll use Caffe's caffe.io.Transformer to do this, but this step is independent of other parts of Caffe, so any custom preprocessing code may be used).\n",
    "\n",
    "Our default CaffeNet is configured to take images in BGR format. Values are expected to start in the range [0, 255] and then have the mean ImageNet pixel value subtracted from them. In addition, the channel dimension is expected as the first (outermost) dimension.\n",
    "\n",
    "As matplotlib will load images with values in the range [0, 1] in RGB format with the channel as the innermost dimension, we are arranging for the needed transformations here."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# set the size of input\n",
    "net.blobs['data'].reshape(50,        # batch size \n",
    "                        3,          # 3-channel(BGR) images\n",
    "                        227, 227)   # image size\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "mu = np.load(caffe_root+'python/caffe/imagenet/ilsvrc_2012_mean.npy') # mu:(3, 256, 256)\n",
    "mu = mu.mean(1).mean(1) # Obtain the mean (BGR) pixel values\n",
    "# print 'mean-substracted values:',zip('BGR',mu)\n",
    "\n",
    "transformer = caffe.io.Transformer({'data':net.blobs['data'].data.shape})\n",
    "\n",
    "transformer.set_transpose('data', (2,0,1))  # move image channels to outermost dimension\n",
    "transformer.set_mean('data', mu)            # subtract the dataset-mean value in each channel\n",
    "transformer.set_raw_scale('data', 255)      # rescale from [0, 1] to [0, 255]\n",
    "transformer.set_channel_swap('data', (2,1,0))  # swap channels from RGB to BGR\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "image = caffe.io.load_image(caffe_root + 'examples/images/cat.jpg')\n",
    "transformed_image = transformer.preprocess('data',image)\n",
    "# plt.imshow(image)\n",
    "# plt.grid('off')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "predicted class is: 0\n"
     ]
    }
   ],
   "source": [
    "net.blobs['data'].data[...] = transformed_image\n",
    "\n",
    "output = net.forward()\n",
    "\n",
    "output_prob = output['prob'][0]\n",
    "\n",
    "print 'predicted class is:',output_prob.argmax()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "output label: n01440764 tench, Tinca tinca\n"
     ]
    }
   ],
   "source": [
    "labels_file = caffe_root + 'data/ilsvrc12/synset_words.txt'\n",
    "labels = np.loadtxt(labels_file, str, delimiter='\\t')\n",
    "print 'output label:',labels[output_prob.argmax()]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "probabilities and labels:\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[(0.99999046, 'n01440764 tench, Tinca tinca'),\n",
       " (9.574238e-06, 'n01443537 goldfish, Carassius auratus')]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_inds = output_prob.argsort()[::-1][:5]\n",
    "\n",
    "print 'probabilities and labels:'\n",
    "zip(output_prob[top_inds], labels[top_inds])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "data\t(50, 3, 227, 227)\n",
      "conv1\t(50, 96, 55, 55)\n",
      "pool1\t(50, 96, 27, 27)\n",
      "norm1\t(50, 96, 27, 27)\n",
      "conv2\t(50, 256, 27, 27)\n",
      "pool2\t(50, 256, 13, 13)\n",
      "norm2\t(50, 256, 13, 13)\n",
      "conv3\t(50, 384, 13, 13)\n",
      "conv4\t(50, 384, 13, 13)\n",
      "conv5\t(50, 256, 13, 13)\n",
      "pool5\t(50, 256, 6, 6)\n",
      "fc6\t(50, 4096)\n",
      "fc7_tune\t(50, 4096)\n",
      "fc8_tune\t(50, 2)\n",
      "prob\t(50, 2)\n"
     ]
    }
   ],
   "source": [
    "for layer_name, blob in net.blobs.iteritems():\n",
    "    print layer_name + '\\t' + str(blob.data.shape)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "conv1\t(96, 3, 11, 11) (96,)\n",
      "conv2\t(256, 48, 5, 5) (256,)\n",
      "conv3\t(384, 256, 3, 3) (384,)\n",
      "conv4\t(384, 192, 3, 3) (384,)\n",
      "conv5\t(256, 192, 3, 3) (256,)\n",
      "fc6\t(4096, 9216) (4096,)\n",
      "fc7_tune\t(4096, 4096) (4096,)\n",
      "fc8_tune\t(2, 4096) (2,)\n"
     ]
    }
   ],
   "source": [
    "for layer_name, param in net.params.iteritems():\n",
    "    print layer_name + '\\t' + str(param[0].data.shape), str(param[1].data.shape)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def vis_square(data):\n",
    "    data = (data - data.min())/(data.max() - data.min())\n",
    "    \n",
    "    n = int(np.ceil(np.sqrt(data.shape[0])))\n",
    "    padding = (((0, n**2 - data.shape[0]), (0, 1), (0, 1)) + ((0, 0),) * (data.ndim - 3))\n",
    "    data = np.pad(data, padding, mode = 'constant', constant_values=1)\n",
    "    \n",
    "    data = data.reshape((n, n) + data.shape[1:]).transpose((0,2,1,3)+tuple(range(4, data.ndim+1)))\n",
    "    data = data.reshape((n * data.shape[1], n * data.shape[3]) + data.shape[4:])\n",
    "    \n",
    "    plt.imshow(data); plt.axis('off')\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# filters = net.params['conv1'][0].data\n",
    "# vis_square(filters.transpose(0, 2, 3, 1))\n",
    "\n",
    "# feat = net.blobs['pool5'].data[0, :]\n",
    "# vis_square(feat)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import scipy.io as scio\n",
    "Data = scio.loadmat('/Users/Bato/Dropbox/tsne_python_5.25/All_PC_DataforBato.mat')\n",
    "I = Data['images']; Ans = Data['state'] # shape of I: (101, 101, 842), shape of Ans:(1, 842)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x10e4426d0>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAloAAAJMCAYAAAAi1AhsAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAGrCRgsAAKAm/wfLhhLOaefUEgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10bf48390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(I[:,:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# This part is done for fine-tuning. Prepare for the txt file needed to write lmdb file.\n",
    "\n",
    "# # def write_image(I):\n",
    "# #     for i in range(0,I.shape[2]):\n",
    "# #         II = (I[:,:,i] - np.mean(I[:,:,i]))/np.mean(I[:,:,i])\n",
    "# #         # II = np.dstack((np.dstack((II,II)),II))\n",
    "# #         plt.imsave('/Users/Bato/Documents/caffe/data/phase_contrast/phase_contrast%s.jpeg' % i,II,cmap='gray')\n",
    "# #     return None\n",
    "# # write_image(I)\n",
    "\n",
    "\n",
    "def write_text(a, b, I):\n",
    "\n",
    "    #a,b is the testing patients number\n",
    "    p = [0, 192, 387, 507, 677, 842]\n",
    "    ar = np.zeros((1, 842))\n",
    "    ar[:,p[a-1]:p[a]] = 1\n",
    "    ar[:,p[b-1]:p[b]] = 1\n",
    "    test_set=[]\n",
    "    train_set=[]\n",
    "    test_index = np.where(ar==1)[1]\n",
    "    train_index = np.where(ar==0)[1]\n",
    "    train_set=open(\"/Users/Bato/Documents/caffe/data/phase_contrast/trainingset.txt\", \"w\")\n",
    "    test_set=open(\"/Users/Bato/Documents/caffe/data/phase_contrast/testingset.txt\", \"w\")\n",
    "    for j in train_index:\n",
    "        train_set.write('/phase_contrast'+str(j)+'.jpeg '+str(int(Ans[:,j]))+'\\n')\n",
    "    train_set.close()\n",
    "    for i in test_index:\n",
    "        test_set.write('/phase_contrast'+str(i)+'.jpeg '+str(int(Ans[:,i]))+'\\n')\n",
    "    \n",
    "    train_images = I[:,:,train_index]\n",
    "    test_images = I[:,:,test_index]\n",
    "    train_labels = Ans[:,train_index]\n",
    "    test_labels = Ans[:,test_index]\n",
    "    \n",
    "    #np.savetxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc8_1500iters/testset_label/'+\n",
    "    #          str(a)+'_'+str(b)+'.txt',test_labels)\n",
    "    \n",
    "    return train_images, test_images, train_labels, test_labels\n",
    "\n",
    "train_I, test_I, train_Ans, test_Ans = write_text(2,4, I)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing the 0 image\n",
      "Processing the 1 image\n",
      "Processing the 2 image\n",
      "Processing the 3 image\n",
      "Processing the 4 image\n",
      "Processing the 5 image\n",
      "Processing the 6 image\n",
      "Processing the 7 image\n",
      "Processing the 8 image\n",
      "Processing the 9 image\n",
      "Processing the 10 image\n",
      "Processing the 11 image\n",
      "Processing the 12 image\n",
      "Processing the 13 image\n",
      "Processing the 14 image\n",
      "Processing the 15 image\n",
      "Processing the 16 image\n",
      "Processing the 17 image\n",
      "Processing the 18 image\n",
      "Processing the 19 image\n",
      "Processing the 20 image\n",
      "Processing the 21 image\n",
      "Processing the 22 image\n",
      "Processing the 23 image\n",
      "Processing the 24 image\n",
      "Processing the 25 image\n",
      "Processing the 26 image\n",
      "Processing the 27 image\n",
      "Processing the 28 image\n",
      "Processing the 29 image\n",
      "Processing the 30 image\n",
      "Processing the 31 image\n",
      "Processing the 32 image\n",
      "Processing the 33 image\n",
      "Processing the 34 image\n",
      "Processing the 35 image\n",
      "Processing the 36 image\n",
      "Processing the 37 image\n",
      "Processing the 38 image\n",
      "Processing the 39 image\n",
      "Processing the 40 image\n",
      "Processing the 41 image\n",
      "Processing the 42 image\n",
      "Processing the 43 image\n",
      "Processing the 44 image\n",
      "Processing the 45 image\n",
      "Processing the 46 image\n",
      "Processing the 47 image\n",
      "Processing the 48 image\n",
      "Processing the 49 image\n",
      "Processing the 50 image\n",
      "Processing the 51 image\n",
      "Processing the 52 image\n",
      "Processing the 53 image\n",
      "Processing the 54 image\n",
      "Processing the 55 image\n",
      "Processing the 56 image\n",
      "Processing the 57 image\n",
      "Processing the 58 image\n",
      "Processing the 59 image\n",
      "Processing the 60 image\n",
      "Processing the 61 image\n",
      "Processing the 62 image\n",
      "Processing the 63 image\n",
      "Processing the 64 image\n",
      "Processing the 65 image\n",
      "Processing the 66 image\n",
      "Processing the 67 image\n",
      "Processing the 68 image\n",
      "Processing the 69 image\n",
      "Processing the 70 image\n",
      "Processing the 71 image\n",
      "Processing the 72 image\n",
      "Processing the 73 image\n",
      "Processing the 74 image\n",
      "Processing the 75 image\n",
      "Processing the 76 image\n",
      "Processing the 77 image\n",
      "Processing the 78 image\n",
      "Processing the 79 image\n",
      "Processing the 80 image\n",
      "Processing the 81 image\n",
      "Processing the 82 image\n",
      "Processing the 83 image\n",
      "Processing the 84 image\n",
      "Processing the 85 image\n",
      "Processing the 86 image\n",
      "Processing the 87 image\n",
      "Processing the 88 image\n",
      "Processing the 89 image\n",
      "Processing the 90 image\n",
      "Processing the 91 image\n",
      "Processing the 92 image\n",
      "Processing the 93 image\n",
      "Processing the 94 image\n",
      "Processing the 95 image\n",
      "Processing the 96 image\n",
      "Processing the 97 image\n",
      "Processing the 98 image\n",
      "Processing the 99 image\n",
      "Processing the 100 image\n",
      "Processing the 101 image\n",
      "Processing the 102 image\n",
      "Processing the 103 image\n",
      "Processing the 104 image\n",
      "Processing the 105 image\n",
      "Processing the 106 image\n",
      "Processing the 107 image\n",
      "Processing the 108 image\n",
      "Processing the 109 image\n",
      "Processing the 110 image\n",
      "Processing the 111 image\n",
      "Processing the 112 image\n",
      "Processing the 113 image\n",
      "Processing the 114 image\n",
      "Processing the 115 image\n",
      "Processing the 116 image\n",
      "Processing the 117 image\n",
      "Processing the 118 image\n",
      "Processing the 119 image\n",
      "Processing the 120 image\n",
      "Processing the 121 image\n",
      "Processing the 122 image\n",
      "Processing the 123 image\n",
      "Processing the 124 image\n",
      "Processing the 125 image\n",
      "Processing the 126 image\n",
      "Processing the 127 image\n",
      "Processing the 128 image\n",
      "Processing the 129 image\n",
      "Processing the 130 image\n",
      "Processing the 131 image\n",
      "Processing the 132 image\n",
      "Processing the 133 image\n",
      "Processing the 134 image\n",
      "Processing the 135 image\n",
      "Processing the 136 image\n",
      "Processing the 137 image\n",
      "Processing the 138 image\n",
      "Processing the 139 image\n",
      "Processing the 140 image\n",
      "Processing the 141 image\n",
      "Processing the 142 image\n",
      "Processing the 143 image\n",
      "Processing the 144 image\n",
      "Processing the 145 image\n",
      "Processing the 146 image\n",
      "Processing the 147 image\n",
      "Processing the 148 image\n",
      "Processing the 149 image\n",
      "Processing the 150 image\n",
      "Processing the 151 image\n",
      "Processing the 152 image\n",
      "Processing the 153 image\n",
      "Processing the 154 image\n",
      "Processing the 155 image\n",
      "Processing the 156 image\n",
      "Processing the 157 image\n",
      "Processing the 158 image\n",
      "Processing the 159 image\n",
      "Processing the 160 image\n",
      "Processing the 161 image\n",
      "Processing the 162 image\n",
      "Processing the 163 image\n",
      "Processing the 164 image\n",
      "Processing the 165 image\n",
      "Processing the 166 image\n",
      "Processing the 167 image\n",
      "Processing the 168 image\n",
      "Processing the 169 image\n",
      "Processing the 170 image\n",
      "Processing the 171 image\n",
      "Processing the 172 image\n",
      "Processing the 173 image\n",
      "Processing the 174 image\n",
      "Processing the 175 image\n",
      "Processing the 176 image\n",
      "Processing the 177 image\n",
      "Processing the 178 image\n",
      "Processing the 179 image\n",
      "Processing the 180 image\n",
      "Processing the 181 image\n",
      "Processing the 182 image\n",
      "Processing the 183 image\n",
      "Processing the 184 image\n",
      "Processing the 185 image\n",
      "Processing the 186 image\n",
      "Processing the 187 image\n",
      "Processing the 188 image\n",
      "Processing the 189 image\n",
      "Processing the 190 image\n",
      "Processing the 191 image\n",
      "Processing the 192 image\n",
      "Processing the 193 image\n",
      "Processing the 194 image\n",
      "Processing the 195 image\n",
      "Processing the 196 image\n",
      "Processing the 197 image\n",
      "Processing the 198 image\n",
      "Processing the 199 image\n",
      "Processing the 200 image\n",
      "Processing the 201 image\n",
      "Processing the 202 image\n",
      "Processing the 203 image\n",
      "Processing the 204 image\n",
      "Processing the 205 image\n",
      "Processing the 206 image\n",
      "Processing the 207 image\n",
      "Processing the 208 image\n",
      "Processing the 209 image\n",
      "Processing the 210 image\n",
      "Processing the 211 image\n",
      "Processing the 212 image\n",
      "Processing the 213 image\n",
      "Processing the 214 image\n",
      "Processing the 215 image\n",
      "Processing the 216 image\n",
      "Processing the 217 image\n",
      "Processing the 218 image\n",
      "Processing the 219 image\n",
      "Processing the 220 image\n",
      "Processing the 221 image\n",
      "Processing the 222 image\n",
      "Processing the 223 image\n",
      "Processing the 224 image\n",
      "Processing the 225 image\n",
      "Processing the 226 image\n",
      "Processing the 227 image\n",
      "Processing the 228 image\n",
      "Processing the 229 image\n",
      "Processing the 230 image\n",
      "Processing the 231 image\n",
      "Processing the 232 image\n",
      "Processing the 233 image\n",
      "Processing the 234 image\n",
      "Processing the 235 image\n",
      "Processing the 236 image\n",
      "Processing the 237 image\n",
      "Processing the 238 image\n",
      "Processing the 239 image\n",
      "Processing the 240 image\n",
      "Processing the 241 image\n",
      "Processing the 242 image\n",
      "Processing the 243 image\n",
      "Processing the 244 image\n",
      "Processing the 245 image\n",
      "Processing the 246 image\n",
      "Processing the 247 image\n",
      "Processing the 248 image\n",
      "Processing the 249 image\n",
      "Processing the 250 image\n",
      "Processing the 251 image\n",
      "Processing the 252 image\n",
      "Processing the 253 image\n",
      "Processing the 254 image\n",
      "Processing the 255 image\n",
      "Processing the 256 image\n",
      "Processing the 257 image\n",
      "Processing the 258 image\n",
      "Processing the 259 image\n",
      "Processing the 260 image\n",
      "Processing the 261 image\n",
      "Processing the 262 image\n",
      "Processing the 263 image\n",
      "Processing the 264 image\n",
      "Processing the 265 image\n",
      "Processing the 266 image\n",
      "Processing the 267 image\n",
      "Processing the 268 image\n",
      "Processing the 269 image\n",
      "Processing the 270 image\n",
      "Processing the 271 image\n",
      "Processing the 272 image\n",
      "Processing the 273 image\n",
      "Processing the 274 image\n",
      "Processing the 275 image\n",
      "Processing the 276 image\n",
      "Processing the 277 image\n",
      "Processing the 278 image\n",
      "Processing the 279 image\n",
      "Processing the 280 image\n",
      "Processing the 281 image\n",
      "Processing the 282 image\n",
      "Processing the 283 image\n",
      "Processing the 284 image\n",
      "Processing the 285 image\n",
      "Processing the 286 image\n",
      "Processing the 287 image\n",
      "Processing the 288 image\n",
      "Processing the 289 image\n",
      "Processing the 290 image\n",
      "Processing the 291 image\n",
      "Processing the 292 image\n",
      "Processing the 293 image\n",
      "Processing the 294 image\n",
      "Processing the 295 image\n",
      "Processing the 296 image\n",
      "Processing the 297 image\n",
      "Processing the 298 image\n",
      "Processing the 299 image\n",
      "Processing the 300 image\n",
      "Processing the 301 image\n",
      "Processing the 302 image\n",
      "Processing the 303 image\n",
      "Processing the 304 image\n",
      "Processing the 305 image\n",
      "Processing the 306 image\n",
      "Processing the 307 image\n",
      "Processing the 308 image\n",
      "Processing the 309 image\n",
      "Processing the 310 image\n",
      "Processing the 311 image\n",
      "Processing the 312 image\n",
      "Processing the 313 image\n",
      "Processing the 314 image\n",
      "Processing the 315 image\n",
      "Processing the 316 image\n",
      "Processing the 317 image\n",
      "Processing the 318 image\n",
      "Processing the 319 image\n",
      "Processing the 320 image\n",
      "Processing the 321 image\n",
      "Processing the 322 image\n",
      "Processing the 323 image\n",
      "Processing the 324 image\n",
      "Processing the 325 image\n",
      "Processing the 326 image\n",
      "Processing the 327 image\n",
      "Processing the 328 image\n",
      "Processing the 329 image\n",
      "Processing the 330 image\n",
      "Processing the 331 image\n",
      "Processing the 332 image\n",
      "Processing the 333 image\n",
      "Processing the 334 image\n",
      "Processing the 335 image\n",
      "Processing the 336 image\n",
      "Processing the 337 image\n",
      "Processing the 338 image\n",
      "Processing the 339 image\n",
      "Processing the 340 image\n",
      "Processing the 341 image\n",
      "Processing the 342 image\n",
      "Processing the 343 image\n",
      "Processing the 344 image\n",
      "Processing the 345 image\n",
      "Processing the 346 image\n",
      "Processing the 347 image\n",
      "Processing the 348 image\n",
      "Processing the 349 image\n",
      "Processing the 350 image\n",
      "Processing the 351 image\n",
      "Processing the 352 image\n",
      "Processing the 353 image\n",
      "Processing the 354 image\n",
      "Processing the 355 image\n",
      "Processing the 356 image\n",
      "Processing the 357 image\n",
      "Processing the 358 image\n",
      "Processing the 359 image\n",
      "Processing the 360 image\n",
      "Processing the 361 image\n",
      "Processing the 362 image\n",
      "Processing the 363 image\n",
      "Processing the 364 image\n"
     ]
    }
   ],
   "source": [
    "def read_med_fine_tuning(images):\n",
    "    \n",
    "    output_matrix = np.zeros([images.shape[2],4096])\n",
    "    for i in range(0,images.shape[2]):\n",
    "        image = images[:,:,i]\n",
    "        image = (image - np.mean(image))/np.mean(image)\n",
    "        image_3d = np.dstack((np.dstack((image,image)),image))\n",
    "        transform_image = transformer.preprocess('data', image_3d)\n",
    "        print 'Processing the {0} image'.format(i)\n",
    "        net.blobs['data'].data[...] = transform_image\n",
    "        output = net.forward()\n",
    "        feature = net.blobs['fc6'].data[0,:]\n",
    "#         output_prob = output['prob'][0]#.argmax()\n",
    "        output_matrix[i,:]= feature#output_prob\n",
    "        \n",
    "    return output_matrix\n",
    "\n",
    "output_matrix = read_med_fine_tuning(test_I)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "np.savetxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc78_1500iters/test2_4/fc7_matrix_2_4.csv',output_matrix)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Data preparation for computing the AUC for fine-tuning\n",
    "from numpy import loadtxt\n",
    "out1_3 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc78_1500iters/test1_3/pro_matrix_1_3.csv')\n",
    "out1_4 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc78_1500iters/test1_4/pro_matrix_1_4.csv')\n",
    "out1_5 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc78_1500iters/test1_5/pro_matrix_1_5.csv')\n",
    "out2_3 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc78_1500iters/test2_3/pro_matrix_2_3.csv')\n",
    "out2_4 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc78_1500iters/test2_4/pro_matrix_2_4.csv')\n",
    "out2_5 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc78_1500iters/test2_5/pro_matrix_2_5.csv')\n",
    "\n",
    "output_matrixs = [out1_3,out1_4,out1_5,out2_3,out2_4,out2_5]\n",
    "\n",
    "Ans1_3 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc8_1500iters/testset_label/1_3.txt')\n",
    "Ans1_4 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc8_1500iters/testset_label/1_4.txt')\n",
    "Ans1_5 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc8_1500iters/testset_label/1_5.txt')\n",
    "Ans2_3 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc8_1500iters/testset_label/2_3.txt')\n",
    "Ans2_4 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc8_1500iters/testset_label/2_4.txt')\n",
    "Ans2_5 = loadtxt('/Users/Bato/Documents/caffe/models/fine_tune_CaffeNet/Tune_fc8_1500iters/testset_label/2_5.txt')\n",
    "\n",
    "ans_matrixs = [Ans1_3, Ans1_4, Ans1_5, Ans2_3, Ans2_4,Ans2_5]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfAAAAFtCAYAAAD4VDh7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x12fd994d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# This part is to computer the AUC when we use fine-tuning.\n",
    "\n",
    "import sklearn.svm as svm\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "import seaborn as sns\n",
    "from scipy import interp\n",
    "\n",
    "name = ['P1&3','P1&4','P1&5','P2&3','P2&4','P2&5']\n",
    "acc = [1.0000,0.9944, 0.7704, 0.5484, 0.6188, 0.7594]\n",
    "acc7 = [1.0000,0.9961, 0.6556, 0.5454, 0.616, 0.7592 ]\n",
    "mean_acc = np.mean(acc)\n",
    "\n",
    "mean_tpr = 0.0\n",
    "mean_fpr = np.linspace(0, 1, 100)\n",
    "all_tpr = []\n",
    "\n",
    "sns.set_palette('muted')\n",
    "palette = np.array(sns.color_palette(\"hls\", 10))\n",
    "\n",
    "mean_tpr = 0.0\n",
    "mean_fpr = np.linspace(0, 1, 100)\n",
    "all_tpr = []\n",
    "\n",
    "for i in range(0, len(output_matrixs)):\n",
    "    fpr, tpr, thresholds = roc_curve(ans_matrixs[i], output_matrixs[i][:,1])\n",
    "    mean_tpr += interp(mean_fpr, fpr, tpr)\n",
    "    mean_tpr[0] = 0.0\n",
    "    roc_auc = auc(fpr, tpr)\n",
    "    plt.plot(fpr, tpr, lw=2,c=palette[i],\n",
    "             label='testing %s,AUC=%0.4f,acc=%0.4f' %(name[i],roc_auc,acc7[i]))\n",
    "    \n",
    "mean_tpr /= len(output_matrixs)\n",
    "mean_tpr[-1] = 1.0\n",
    "mean_auc = auc(mean_fpr, mean_tpr)\n",
    "\n",
    "plt.xlim([-0.05, 1.05])\n",
    "plt.ylim([-0.05, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC curve of fine-tuning fc8')\n",
    "plt.legend(loc=4)\n",
    "plt.grid('off')\n",
    "plt.plot(mean_fpr, mean_tpr, 'k--',\n",
    "         label='Mean ROC (AUC = %0.4f),Mean acc=%0.4f' % (mean_auc,mean_acc), lw=2)\n",
    "plt.legend(loc=4)\n",
    "plt.plot([0, 1], [0, 1], '--', color=(0.6, 0.6, 0.6), label='Luck')\n",
    "\n",
    "plt.savefig('/Users/Bato/Dropbox/tsne_python_5.25/Feature/CaffeNet/fine-tuning/fc78/AUC.png')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing the 0 image\n",
      "Processing the 1 image\n",
      "Processing the 2 image\n",
      "Processing the 3 image\n",
      "Processing the 4 image\n",
      "Processing the 5 image\n",
      "Processing the 6 image\n",
      "Processing the 7 image\n",
      "Processing the 8 image\n",
      "Processing the 9 image\n",
      "Processing the 10 image\n",
      "Processing the 11 image\n",
      "Processing the 12 image\n",
      "Processing the 13 image\n",
      "Processing the 14 image\n",
      "Processing the 15 image\n",
      "Processing the 16 image\n",
      "Processing the 17 image\n",
      "Processing the 18 image\n",
      "Processing the 19 image\n",
      "Processing the 20 image\n",
      "Processing the 21 image\n",
      "Processing the 22 image\n",
      "Processing the 23 image\n",
      "Processing the 24 image\n",
      "Processing the 25 image\n",
      "Processing the 26 image\n",
      "Processing the 27 image\n",
      "Processing the 28 image\n",
      "Processing the 29 image\n",
      "Processing the 30 image\n",
      "Processing the 31 image\n",
      "Processing the 32 image\n",
      "Processing the 33 image\n",
      "Processing the 34 image\n",
      "Processing the 35 image\n",
      "Processing the 36 image\n",
      "Processing the 37 image\n",
      "Processing the 38 image\n",
      "Processing the 39 image\n",
      "Processing the 40 image\n",
      "Processing the 41 image\n",
      "Processing the 42 image\n",
      "Processing the 43 image\n",
      "Processing the 44 image\n",
      "Processing the 45 image\n",
      "Processing the 46 image\n",
      "Processing the 47 image\n",
      "Processing the 48 image\n",
      "Processing the 49 image\n",
      "Processing the 50 image\n",
      "Processing the 51 image\n",
      "Processing the 52 image\n",
      "Processing the 53 image\n",
      "Processing the 54 image\n",
      "Processing the 55 image\n",
      "Processing the 56 image\n",
      "Processing the 57 image\n",
      "Processing the 58 image\n",
      "Processing the 59 image\n",
      "Processing the 60 image\n",
      "Processing the 61 image\n",
      "Processing the 62 image\n",
      "Processing the 63 image\n",
      "Processing the 64 image\n",
      "Processing the 65 image\n",
      "Processing the 66 image\n",
      "Processing the 67 image\n",
      "Processing the 68 image\n",
      "Processing the 69 image\n",
      "Processing the 70 image\n",
      "Processing the 71 image\n",
      "Processing the 72 image\n",
      "Processing the 73 image\n",
      "Processing the 74 image\n",
      "Processing the 75 image\n",
      "Processing the 76 image\n",
      "Processing the 77 image\n",
      "Processing the 78 image\n",
      "Processing the 79 image\n",
      "Processing the 80 image\n",
      "Processing the 81 image\n",
      "Processing the 82 image\n",
      "Processing the 83 image\n",
      "Processing the 84 image\n",
      "Processing the 85 image\n",
      "Processing the 86 image\n",
      "Processing the 87 image\n",
      "Processing the 88 image\n",
      "Processing the 89 image\n",
      "Processing the 90 image\n",
      "Processing the 91 image\n",
      "Processing the 92 image\n",
      "Processing the 93 image\n",
      "Processing the 94 image\n",
      "Processing the 95 image\n",
      "Processing the 96 image\n",
      "Processing the 97 image\n",
      "Processing the 98 image\n",
      "Processing the 99 image\n",
      "Processing the 100 image\n",
      "Processing the 101 image\n",
      "Processing the 102 image\n",
      "Processing the 103 image\n",
      "Processing the 104 image\n",
      "Processing the 105 image\n",
      "Processing the 106 image\n",
      "Processing the 107 image\n",
      "Processing the 108 image\n",
      "Processing the 109 image\n",
      "Processing the 110 image\n",
      "Processing the 111 image\n",
      "Processing the 112 image\n",
      "Processing the 113 image\n",
      "Processing the 114 image\n",
      "Processing the 115 image\n",
      "Processing the 116 image\n",
      "Processing the 117 image\n",
      "Processing the 118 image\n",
      "Processing the 119 image\n",
      "Processing the 120 image\n",
      "Processing the 121 image\n",
      "Processing the 122 image\n",
      "Processing the 123 image\n",
      "Processing the 124 image\n",
      "Processing the 125 image\n",
      "Processing the 126 image\n",
      "Processing the 127 image\n",
      "Processing the 128 image\n",
      "Processing the 129 image\n",
      "Processing the 130 image\n",
      "Processing the 131 image\n",
      "Processing the 132 image\n",
      "Processing the 133 image\n",
      "Processing the 134 image\n",
      "Processing the 135 image\n",
      "Processing the 136 image\n",
      "Processing the 137 image\n",
      "Processing the 138 image\n",
      "Processing the 139 image\n",
      "Processing the 140 image\n",
      "Processing the 141 image\n",
      "Processing the 142 image\n",
      "Processing the 143 image\n",
      "Processing the 144 image\n",
      "Processing the 145 image\n",
      "Processing the 146 image\n",
      "Processing the 147 image\n",
      "Processing the 148 image\n",
      "Processing the 149 image\n",
      "Processing the 150 image\n",
      "Processing the 151 image\n",
      "Processing the 152 image\n",
      "Processing the 153 image\n",
      "Processing the 154 image\n",
      "Processing the 155 image\n",
      "Processing the 156 image\n",
      "Processing the 157 image\n",
      "Processing the 158 image\n",
      "Processing the 159 image\n",
      "Processing the 160 image\n",
      "Processing the 161 image\n",
      "Processing the 162 image\n",
      "Processing the 163 image\n",
      "Processing the 164 image\n",
      "Processing the 165 image\n",
      "Processing the 166 image\n",
      "Processing the 167 image\n",
      "Processing the 168 image\n",
      "Processing the 169 image\n",
      "Processing the 170 image\n",
      "Processing the 171 image\n",
      "Processing the 172 image\n",
      "Processing the 173 image\n",
      "Processing the 174 image\n",
      "Processing the 175 image\n",
      "Processing the 176 image\n",
      "Processing the 177 image\n",
      "Processing the 178 image\n",
      "Processing the 179 image\n",
      "Processing the 180 image\n",
      "Processing the 181 image\n",
      "Processing the 182 image\n",
      "Processing the 183 image\n",
      "Processing the 184 image\n",
      "Processing the 185 image\n",
      "Processing the 186 image\n",
      "Processing the 187 image\n",
      "Processing the 188 image\n",
      "Processing the 189 image\n",
      "Processing the 190 image\n",
      "Processing the 191 image\n",
      "Processing the 192 image\n",
      "Processing the 193 image\n",
      "Processing the 194 image\n",
      "Processing the 195 image\n",
      "Processing the 196 image\n",
      "Processing the 197 image\n",
      "Processing the 198 image\n",
      "Processing the 199 image\n",
      "Processing the 200 image\n",
      "Processing the 201 image\n",
      "Processing the 202 image\n",
      "Processing the 203 image\n",
      "Processing the 204 image\n",
      "Processing the 205 image\n",
      "Processing the 206 image\n",
      "Processing the 207 image\n",
      "Processing the 208 image\n",
      "Processing the 209 image\n",
      "Processing the 210 image\n",
      "Processing the 211 image\n",
      "Processing the 212 image\n",
      "Processing the 213 image\n",
      "Processing the 214 image\n",
      "Processing the 215 image\n",
      "Processing the 216 image\n",
      "Processing the 217 image\n",
      "Processing the 218 image\n",
      "Processing the 219 image\n",
      "Processing the 220 image\n",
      "Processing the 221 image\n",
      "Processing the 222 image\n",
      "Processing the 223 image\n",
      "Processing the 224 image\n",
      "Processing the 225 image\n",
      "Processing the 226 image\n",
      "Processing the 227 image\n",
      "Processing the 228 image\n",
      "Processing the 229 image\n",
      "Processing the 230 image\n",
      "Processing the 231 image\n",
      "Processing the 232 image\n",
      "Processing the 233 image\n",
      "Processing the 234 image\n",
      "Processing the 235 image\n",
      "Processing the 236 image\n",
      "Processing the 237 image\n",
      "Processing the 238 image\n",
      "Processing the 239 image\n",
      "Processing the 240 image\n",
      "Processing the 241 image\n",
      "Processing the 242 image\n",
      "Processing the 243 image\n",
      "Processing the 244 image\n",
      "Processing the 245 image\n",
      "Processing the 246 image\n",
      "Processing the 247 image\n",
      "Processing the 248 image\n",
      "Processing the 249 image\n",
      "Processing the 250 image\n",
      "Processing the 251 image\n",
      "Processing the 252 image\n",
      "Processing the 253 image\n",
      "Processing the 254 image\n",
      "Processing the 255 image\n",
      "Processing the 256 image\n",
      "Processing the 257 image\n",
      "Processing the 258 image\n",
      "Processing the 259 image\n",
      "Processing the 260 image\n",
      "Processing the 261 image\n",
      "Processing the 262 image\n",
      "Processing the 263 image\n",
      "Processing the 264 image\n",
      "Processing the 265 image\n",
      "Processing the 266 image\n",
      "Processing the 267 image\n",
      "Processing the 268 image\n",
      "Processing the 269 image\n",
      "Processing the 270 image\n",
      "Processing the 271 image\n",
      "Processing the 272 image\n",
      "Processing the 273 image\n",
      "Processing the 274 image\n",
      "Processing the 275 image\n",
      "Processing the 276 image\n",
      "Processing the 277 image\n",
      "Processing the 278 image\n",
      "Processing the 279 image\n",
      "Processing the 280 image\n",
      "Processing the 281 image\n",
      "Processing the 282 image\n",
      "Processing the 283 image\n",
      "Processing the 284 image\n",
      "Processing the 285 image\n",
      "Processing the 286 image\n",
      "Processing the 287 image\n",
      "Processing the 288 image\n",
      "Processing the 289 image\n",
      "Processing the 290 image\n",
      "Processing the 291 image\n",
      "Processing the 292 image\n",
      "Processing the 293 image\n",
      "Processing the 294 image\n",
      "Processing the 295 image\n",
      "Processing the 296 image\n",
      "Processing the 297 image\n",
      "Processing the 298 image\n",
      "Processing the 299 image\n",
      "Processing the 300 image\n",
      "Processing the 301 image\n",
      "Processing the 302 image\n",
      "Processing the 303 image\n",
      "Processing the 304 image\n",
      "Processing the 305 image\n",
      "Processing the 306 image\n",
      "Processing the 307 image\n",
      "Processing the 308 image\n",
      "Processing the 309 image\n",
      "Processing the 310 image\n",
      "Processing the 311 image\n",
      "Processing the 312 image\n",
      "Processing the 313 image\n",
      "Processing the 314 image\n",
      "Processing the 315 image\n",
      "Processing the 316 image\n",
      "Processing the 317 image\n",
      "Processing the 318 image\n",
      "Processing the 319 image\n",
      "Processing the 320 image\n",
      "Processing the 321 image\n",
      "Processing the 322 image\n",
      "Processing the 323 image\n",
      "Processing the 324 image\n",
      "Processing the 325 image\n",
      "Processing the 326 image\n",
      "Processing the 327 image\n",
      "Processing the 328 image\n",
      "Processing the 329 image\n",
      "Processing the 330 image\n",
      "Processing the 331 image\n",
      "Processing the 332 image\n",
      "Processing the 333 image\n",
      "Processing the 334 image\n",
      "Processing the 335 image\n",
      "Processing the 336 image\n",
      "Processing the 337 image\n",
      "Processing the 338 image\n",
      "Processing the 339 image\n",
      "Processing the 340 image\n",
      "Processing the 341 image\n",
      "Processing the 342 image\n",
      "Processing the 343 image\n",
      "Processing the 344 image\n",
      "Processing the 345 image\n",
      "Processing the 346 image\n",
      "Processing the 347 image\n",
      "Processing the 348 image\n",
      "Processing the 349 image\n",
      "Processing the 350 image\n",
      "Processing the 351 image\n",
      "Processing the 352 image\n",
      "Processing the 353 image\n",
      "Processing the 354 image\n",
      "Processing the 355 image\n",
      "Processing the 356 image\n",
      "Processing the 357 image\n",
      "Processing the 358 image\n",
      "Processing the 359 image\n",
      "Processing the 360 image\n",
      "Processing the 361 image\n",
      "Processing the 362 image\n",
      "Processing the 363 image\n",
      "Processing the 364 image\n",
      "Processing the 365 image\n",
      "Processing the 366 image\n",
      "Processing the 367 image\n",
      "Processing the 368 image\n",
      "Processing the 369 image\n",
      "Processing the 370 image\n",
      "Processing the 371 image\n",
      "Processing the 372 image\n",
      "Processing the 373 image\n",
      "Processing the 374 image\n",
      "Processing the 375 image\n",
      "Processing the 376 image\n",
      "Processing the 377 image\n",
      "Processing the 378 image\n",
      "Processing the 379 image\n",
      "Processing the 380 image\n",
      "Processing the 381 image\n",
      "Processing the 382 image\n",
      "Processing the 383 image\n",
      "Processing the 384 image\n",
      "Processing the 385 image\n",
      "Processing the 386 image\n",
      "Processing the 387 image\n",
      "Processing the 388 image\n",
      "Processing the 389 image\n",
      "Processing the 390 image\n",
      "Processing the 391 image\n",
      "Processing the 392 image\n",
      "Processing the 393 image\n",
      "Processing the 394 image\n",
      "Processing the 395 image\n",
      "Processing the 396 image\n",
      "Processing the 397 image\n",
      "Processing the 398 image\n",
      "Processing the 399 image\n",
      "Processing the 400 image\n",
      "Processing the 401 image\n",
      "Processing the 402 image\n",
      "Processing the 403 image\n",
      "Processing the 404 image\n",
      "Processing the 405 image\n",
      "Processing the 406 image\n",
      "Processing the 407 image\n",
      "Processing the 408 image\n",
      "Processing the 409 image\n",
      "Processing the 410 image\n",
      "Processing the 411 image\n",
      "Processing the 412 image\n",
      "Processing the 413 image\n",
      "Processing the 414 image\n",
      "Processing the 415 image\n",
      "Processing the 416 image\n",
      "Processing the 417 image\n",
      "Processing the 418 image\n",
      "Processing the 419 image\n",
      "Processing the 420 image\n",
      "Processing the 421 image\n",
      "Processing the 422 image\n",
      "Processing the 423 image\n",
      "Processing the 424 image\n",
      "Processing the 425 image\n",
      "Processing the 426 image\n",
      "Processing the 427 image\n",
      "Processing the 428 image\n",
      "Processing the 429 image\n",
      "Processing the 430 image\n",
      "Processing the 431 image\n",
      "Processing the 432 image\n",
      "Processing the 433 image\n",
      "Processing the 434 image\n",
      "Processing the 435 image\n",
      "Processing the 436 image\n",
      "Processing the 437 image\n",
      "Processing the 438 image\n",
      "Processing the 439 image\n",
      "Processing the 440 image\n",
      "Processing the 441 image\n",
      "Processing the 442 image\n",
      "Processing the 443 image\n",
      "Processing the 444 image\n",
      "Processing the 445 image\n",
      "Processing the 446 image\n",
      "Processing the 447 image\n",
      "Processing the 448 image\n",
      "Processing the 449 image\n",
      "Processing the 450 image\n",
      "Processing the 451 image\n",
      "Processing the 452 image\n",
      "Processing the 453 image\n",
      "Processing the 454 image\n",
      "Processing the 455 image\n",
      "Processing the 456 image\n",
      "Processing the 457 image\n",
      "Processing the 458 image\n",
      "Processing the 459 image\n",
      "Processing the 460 image\n",
      "Processing the 461 image\n",
      "Processing the 462 image\n",
      "Processing the 463 image\n",
      "Processing the 464 image\n",
      "Processing the 465 image\n",
      "Processing the 466 image\n",
      "Processing the 467 image\n",
      "Processing the 468 image\n",
      "Processing the 469 image\n",
      "Processing the 470 image\n",
      "Processing the 471 image\n",
      "Processing the 472 image\n",
      "Processing the 473 image\n",
      "Processing the 474 image\n",
      "Processing the 475 image\n",
      "Processing the 476 image\n",
      "Processing the 477 image\n",
      "Processing the 478 image\n",
      "Processing the 479 image\n",
      "Processing the 480 image\n",
      "Processing the 481 image\n",
      "Processing the 482 image\n",
      "Processing the 483 image\n",
      "Processing the 484 image\n",
      "Processing the 485 image\n",
      "Processing the 486 image\n",
      "Processing the 487 image\n",
      "Processing the 488 image\n",
      "Processing the 489 image\n",
      "Processing the 490 image\n",
      "Processing the 491 image\n",
      "Processing the 492 image\n",
      "Processing the 493 image\n",
      "Processing the 494 image\n",
      "Processing the 495 image\n",
      "Processing the 496 image\n",
      "Processing the 497 image\n",
      "Processing the 498 image\n",
      "Processing the 499 image\n",
      "Processing the 500 image\n",
      "Processing the 501 image\n",
      "Processing the 502 image\n",
      "Processing the 503 image\n",
      "Processing the 504 image\n",
      "Processing the 505 image\n",
      "Processing the 506 image\n",
      "Processing the 507 image\n",
      "Processing the 508 image\n",
      "Processing the 509 image\n",
      "Processing the 510 image\n",
      "Processing the 511 image\n",
      "Processing the 512 image\n",
      "Processing the 513 image\n",
      "Processing the 514 image\n",
      "Processing the 515 image\n",
      "Processing the 516 image\n",
      "Processing the 517 image\n",
      "Processing the 518 image\n",
      "Processing the 519 image\n",
      "Processing the 520 image\n",
      "Processing the 521 image\n",
      "Processing the 522 image\n",
      "Processing the 523 image\n",
      "Processing the 524 image\n",
      "Processing the 525 image\n",
      "Processing the 526 image\n",
      "Processing the 527 image\n",
      "Processing the 528 image\n",
      "Processing the 529 image\n",
      "Processing the 530 image\n",
      "Processing the 531 image\n",
      "Processing the 532 image\n",
      "Processing the 533 image\n",
      "Processing the 534 image\n",
      "Processing the 535 image\n",
      "Processing the 536 image\n",
      "Processing the 537 image\n",
      "Processing the 538 image\n",
      "Processing the 539 image\n",
      "Processing the 540 image\n",
      "Processing the 541 image\n",
      "Processing the 542 image\n",
      "Processing the 543 image\n",
      "Processing the 544 image\n",
      "Processing the 545 image\n",
      "Processing the 546 image\n",
      "Processing the 547 image\n",
      "Processing the 548 image\n",
      "Processing the 549 image\n",
      "Processing the 550 image\n",
      "Processing the 551 image\n",
      "Processing the 552 image\n",
      "Processing the 553 image\n",
      "Processing the 554 image\n",
      "Processing the 555 image\n",
      "Processing the 556 image\n",
      "Processing the 557 image\n",
      "Processing the 558 image\n",
      "Processing the 559 image\n",
      "Processing the 560 image\n",
      "Processing the 561 image\n",
      "Processing the 562 image\n",
      "Processing the 563 image\n",
      "Processing the 564 image\n",
      "Processing the 565 image\n",
      "Processing the 566 image\n",
      "Processing the 567 image\n",
      "Processing the 568 image\n",
      "Processing the 569 image\n",
      "Processing the 570 image\n",
      "Processing the 571 image\n",
      "Processing the 572 image\n",
      "Processing the 573 image\n",
      "Processing the 574 image\n",
      "Processing the 575 image\n",
      "Processing the 576 image\n",
      "Processing the 577 image\n",
      "Processing the 578 image\n",
      "Processing the 579 image\n",
      "Processing the 580 image\n",
      "Processing the 581 image\n",
      "Processing the 582 image\n",
      "Processing the 583 image\n",
      "Processing the 584 image\n",
      "Processing the 585 image\n",
      "Processing the 586 image\n",
      "Processing the 587 image\n",
      "Processing the 588 image\n",
      "Processing the 589 image\n",
      "Processing the 590 image\n",
      "Processing the 591 image\n",
      "Processing the 592 image\n",
      "Processing the 593 image\n",
      "Processing the 594 image\n",
      "Processing the 595 image\n",
      "Processing the 596 image\n",
      "Processing the 597 image\n",
      "Processing the 598 image\n",
      "Processing the 599 image\n",
      "Processing the 600 image\n",
      "Processing the 601 image\n",
      "Processing the 602 image\n",
      "Processing the 603 image\n",
      "Processing the 604 image\n",
      "Processing the 605 image\n",
      "Processing the 606 image\n",
      "Processing the 607 image\n",
      "Processing the 608 image\n",
      "Processing the 609 image\n",
      "Processing the 610 image\n",
      "Processing the 611 image\n",
      "Processing the 612 image\n",
      "Processing the 613 image\n",
      "Processing the 614 image\n",
      "Processing the 615 image\n",
      "Processing the 616 image\n",
      "Processing the 617 image\n",
      "Processing the 618 image\n",
      "Processing the 619 image\n",
      "Processing the 620 image\n",
      "Processing the 621 image\n",
      "Processing the 622 image\n",
      "Processing the 623 image\n",
      "Processing the 624 image\n",
      "Processing the 625 image\n",
      "Processing the 626 image\n",
      "Processing the 627 image\n",
      "Processing the 628 image\n",
      "Processing the 629 image\n",
      "Processing the 630 image\n",
      "Processing the 631 image\n",
      "Processing the 632 image\n",
      "Processing the 633 image\n",
      "Processing the 634 image\n",
      "Processing the 635 image\n",
      "Processing the 636 image\n",
      "Processing the 637 image\n",
      "Processing the 638 image\n",
      "Processing the 639 image\n",
      "Processing the 640 image\n",
      "Processing the 641 image\n",
      "Processing the 642 image\n",
      "Processing the 643 image\n",
      "Processing the 644 image\n",
      "Processing the 645 image\n",
      "Processing the 646 image\n",
      "Processing the 647 image\n",
      "Processing the 648 image\n",
      "Processing the 649 image\n",
      "Processing the 650 image\n",
      "Processing the 651 image\n",
      "Processing the 652 image\n",
      "Processing the 653 image\n",
      "Processing the 654 image\n",
      "Processing the 655 image\n",
      "Processing the 656 image\n",
      "Processing the 657 image\n",
      "Processing the 658 image\n",
      "Processing the 659 image\n",
      "Processing the 660 image\n",
      "Processing the 661 image\n",
      "Processing the 662 image\n",
      "Processing the 663 image\n",
      "Processing the 664 image\n",
      "Processing the 665 image\n",
      "Processing the 666 image\n",
      "Processing the 667 image\n",
      "Processing the 668 image\n",
      "Processing the 669 image\n",
      "Processing the 670 image\n",
      "Processing the 671 image\n",
      "Processing the 672 image\n",
      "Processing the 673 image\n",
      "Processing the 674 image\n",
      "Processing the 675 image\n",
      "Processing the 676 image\n",
      "Processing the 677 image\n",
      "Processing the 678 image\n",
      "Processing the 679 image\n",
      "Processing the 680 image\n",
      "Processing the 681 image\n",
      "Processing the 682 image\n",
      "Processing the 683 image\n",
      "Processing the 684 image\n",
      "Processing the 685 image\n",
      "Processing the 686 image\n",
      "Processing the 687 image\n",
      "Processing the 688 image\n",
      "Processing the 689 image\n",
      "Processing the 690 image\n",
      "Processing the 691 image\n",
      "Processing the 692 image\n",
      "Processing the 693 image\n",
      "Processing the 694 image\n",
      "Processing the 695 image\n",
      "Processing the 696 image\n",
      "Processing the 697 image\n",
      "Processing the 698 image\n",
      "Processing the 699 image\n",
      "Processing the 700 image\n",
      "Processing the 701 image\n",
      "Processing the 702 image\n",
      "Processing the 703 image\n",
      "Processing the 704 image\n",
      "Processing the 705 image\n",
      "Processing the 706 image\n",
      "Processing the 707 image\n",
      "Processing the 708 image\n",
      "Processing the 709 image\n",
      "Processing the 710 image\n",
      "Processing the 711 image\n",
      "Processing the 712 image\n",
      "Processing the 713 image\n",
      "Processing the 714 image\n",
      "Processing the 715 image\n",
      "Processing the 716 image\n",
      "Processing the 717 image\n",
      "Processing the 718 image\n",
      "Processing the 719 image\n",
      "Processing the 720 image\n",
      "Processing the 721 image\n",
      "Processing the 722 image\n",
      "Processing the 723 image\n",
      "Processing the 724 image\n",
      "Processing the 725 image\n",
      "Processing the 726 image\n",
      "Processing the 727 image\n",
      "Processing the 728 image\n",
      "Processing the 729 image\n",
      "Processing the 730 image\n",
      "Processing the 731 image\n",
      "Processing the 732 image\n",
      "Processing the 733 image\n",
      "Processing the 734 image\n",
      "Processing the 735 image\n",
      "Processing the 736 image\n",
      "Processing the 737 image\n",
      "Processing the 738 image\n",
      "Processing the 739 image\n",
      "Processing the 740 image\n",
      "Processing the 741 image\n",
      "Processing the 742 image\n",
      "Processing the 743 image\n",
      "Processing the 744 image\n",
      "Processing the 745 image\n",
      "Processing the 746 image\n",
      "Processing the 747 image\n",
      "Processing the 748 image\n",
      "Processing the 749 image\n",
      "Processing the 750 image\n",
      "Processing the 751 image\n",
      "Processing the 752 image\n",
      "Processing the 753 image\n",
      "Processing the 754 image\n",
      "Processing the 755 image\n",
      "Processing the 756 image\n",
      "Processing the 757 image\n",
      "Processing the 758 image\n",
      "Processing the 759 image\n",
      "Processing the 760 image\n",
      "Processing the 761 image\n",
      "Processing the 762 image\n",
      "Processing the 763 image\n",
      "Processing the 764 image\n",
      "Processing the 765 image\n",
      "Processing the 766 image\n",
      "Processing the 767 image\n",
      "Processing the 768 image\n",
      "Processing the 769 image\n",
      "Processing the 770 image\n",
      "Processing the 771 image\n",
      "Processing the 772 image\n",
      "Processing the 773 image\n",
      "Processing the 774 image\n",
      "Processing the 775 image\n",
      "Processing the 776 image\n",
      "Processing the 777 image\n",
      "Processing the 778 image\n",
      "Processing the 779 image\n",
      "Processing the 780 image\n",
      "Processing the 781 image\n",
      "Processing the 782 image\n",
      "Processing the 783 image\n",
      "Processing the 784 image\n",
      "Processing the 785 image\n",
      "Processing the 786 image\n",
      "Processing the 787 image\n",
      "Processing the 788 image\n",
      "Processing the 789 image\n",
      "Processing the 790 image\n",
      "Processing the 791 image\n",
      "Processing the 792 image\n",
      "Processing the 793 image\n",
      "Processing the 794 image\n",
      "Processing the 795 image\n",
      "Processing the 796 image\n",
      "Processing the 797 image\n",
      "Processing the 798 image\n",
      "Processing the 799 image\n",
      "Processing the 800 image\n",
      "Processing the 801 image\n",
      "Processing the 802 image\n",
      "Processing the 803 image\n",
      "Processing the 804 image\n",
      "Processing the 805 image\n",
      "Processing the 806 image\n",
      "Processing the 807 image\n",
      "Processing the 808 image\n",
      "Processing the 809 image\n",
      "Processing the 810 image\n",
      "Processing the 811 image\n",
      "Processing the 812 image\n",
      "Processing the 813 image\n",
      "Processing the 814 image\n",
      "Processing the 815 image\n",
      "Processing the 816 image\n",
      "Processing the 817 image\n",
      "Processing the 818 image\n",
      "Processing the 819 image\n",
      "Processing the 820 image\n",
      "Processing the 821 image\n",
      "Processing the 822 image\n",
      "Processing the 823 image\n",
      "Processing the 824 image\n",
      "Processing the 825 image\n",
      "Processing the 826 image\n",
      "Processing the 827 image\n",
      "Processing the 828 image\n",
      "Processing the 829 image\n",
      "Processing the 830 image\n",
      "Processing the 831 image\n",
      "Processing the 832 image\n",
      "Processing the 833 image\n",
      "Processing the 834 image\n",
      "Processing the 835 image\n",
      "Processing the 836 image\n",
      "Processing the 837 image\n",
      "Processing the 838 image\n",
      "Processing the 839 image\n",
      "Processing the 840 image\n",
      "Processing the 841 image\n"
     ]
    }
   ],
   "source": [
    "# This part is to generate off-the-shell CNN features. \n",
    "\n",
    "def read_med_off_the_shell(images):\n",
    "    \n",
    "    feat_matrix = np.zeros((images.shape[2],4096))\n",
    "    for i in range(0, images.shape[2]):\n",
    "        image = images[:,:,i]\n",
    "        image = (image - np.mean(image))/np.mean(image)\n",
    "        image_3d = np.dstack((np.dstack((image,image)),image))\n",
    "        transform_image = transformer.preprocess('data', image_3d)\n",
    "        print 'Processing the {0} image'.format(i)\n",
    "        net.blobs['data'].data[...] = transform_image\n",
    "        output = net.forward()\n",
    "        feature = net.blobs['fc6'].data[0,:]\n",
    "        # feat_matrix[i, :] = np.mean(np.mean(feature, axis=1),axis=1)\n",
    "        feat_matrix[i,:] = feature\n",
    "    \n",
    "    return feat_matrix\n",
    "\n",
    "feature_matrix = read_med_off_the_shell(I)\n",
    "np.savetxt('/Users/Bato/Dropbox/tsne_python_5.25/Feature/CaffeNet/fine-tuning/OFF-the-shell/test1_3_fc6.csv',feature_matrix,delimiter=\",\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Split the whole feature matrix into training/testing sets patient-wise. \n",
    "feature_matrix =np.loadtxt('/Users/Bato/Dropbox/tsne_python_5.25/Feature/CaffeNet/fine-tuning/OFF-the-shell/test1_3/test1_3_fc6.csv',delimiter=\",\")\n",
    "# feature_matrix =np.loadtxt('/Users/Bato/Dropbox/tsne_python_5.25/Feature/inception3/off-the-shelf/mix1/mix1.csv',delimiter=\",\")\n",
    "\n",
    "def split_matrix_to_patient(feature_matrix):\n",
    "    \n",
    "    p1 = np.concatenate((feature_matrix[0:192,:], Ans.T[0:192]),axis=1)\n",
    "    p2 = np.concatenate((feature_matrix[192:387,:], Ans.T[192:387]),axis=1)\n",
    "    p3 = np.concatenate((feature_matrix[387:507,:], Ans.T[387:507]),axis=1)\n",
    "    p4 = np.concatenate((feature_matrix[507:677,:], Ans.T[507:677]),axis=1)\n",
    "    p5 = np.concatenate((feature_matrix[677:,:], Ans.T[677:]),axis=1)\n",
    "\n",
    "    tr1 = np.concatenate((p2,p4,p5), axis=0)\n",
    "    tr2 = np.concatenate((p2,p3,p5), axis=0)\n",
    "    tr3 = np.concatenate((p2,p3,p4), axis=0)\n",
    "    tr4 = np.concatenate((p1,p4,p5), axis=0)\n",
    "    tr5 = np.concatenate((p1,p3,p5), axis=0)\n",
    "    tr6 = np.concatenate((p1,p3,p4), axis=0)\n",
    "\n",
    "    t1 = np.concatenate((p1,p3),axis=0)\n",
    "    t2 = np.concatenate((p1,p4),axis=0)\n",
    "    t3 = np.concatenate((p1,p5),axis=0)\n",
    "    t4 = np.concatenate((p2,p3),axis=0)\n",
    "    t5 = np.concatenate((p2,p4),axis=0)\n",
    "    t6 = np.concatenate((p2,p5),axis=0)\n",
    "\n",
    "    return tr1, tr2, tr3, tr4, tr5, tr6, t1, t2, t3, t4, t5, t6\n",
    "\n",
    "tr1, tr2, tr3, tr4, tr5, tr6, t1, t2, t3, t4, t5, t6 = split_matrix_to_patient(feature_matrix)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false,
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmkAAAIICAYAAAA12XktAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nn9+b6wAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x119cc20d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# This part is to compute the AUC when we use off-the-shell CNN features to classify.\n",
    "# This features are all high-dimensional features. \n",
    "\n",
    "import sklearn.svm as svm\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "import seaborn as sns\n",
    "from scipy import interp\n",
    "\n",
    "sns.set_palette('muted')\n",
    "palette = np.array(sns.color_palette(\"hls\", 10))\n",
    "\n",
    "random_state = np.random.RandomState(0)\n",
    "clf = svm.SVC(gamma=0.001, C=100., probability=True, random_state=random_state)\n",
    "train = [tr1,tr2,tr3,tr4,tr5,tr6]\n",
    "test = [t1,t2,t3,t4,t5,t6]\n",
    "name = ['P1&3','P1&4','P1&5','P2&3','P2&4','P2&5']\n",
    "\n",
    "mean_tpr3 = 0.0\n",
    "mean_fpr3 = np.linspace(0, 1, 100)\n",
    "mean_accuracy = []\n",
    "std_AUC = []\n",
    "\n",
    "for i,val in enumerate(train):\n",
    "    \n",
    "    clf.fit(train[i][:,:-1],np.ravel(train[i][:,-1:]))\n",
    "\n",
    "    probas = clf.predict_proba(test[i][:,:-1])\n",
    "    fpr, tpr, thresholds = roc_curve(test[i][:,-1:], probas[:,1])\n",
    "    mean_tpr3 += interp(mean_fpr3, fpr, tpr)\n",
    "    mean_tpr3[0] = 0.0\n",
    "    roc_auc = auc(fpr, tpr)\n",
    "    std_AUC.append(roc_auc)\n",
    "    auccracy = clf.score(test[i][:,:-1], test[i][:,-1:])\n",
    "    mean_accuracy.append(auccracy)\n",
    "    plt.plot(fpr, tpr, lw=2,c=palette[i],\n",
    "             label='testing %s,AUC=%0.4f,acc=%0.4f' % (name[i],roc_auc,auccracy))\n",
    "\n",
    "mpl.rc('figure',figsize=(15,8))\n",
    "mean_tpr3 /= len(train)\n",
    "mean_tpr3[-1] = 1.0\n",
    "mean_auc3 = auc(mean_fpr3, mean_tpr3)\n",
    "std_accuracy = np.std(mean_accuracy)\n",
    "std_AUC = np.std(std_AUC)\n",
    "mean_accuracy = np.sum(mean_accuracy)/len(mean_accuracy)\n",
    "# plt.plot(mean_fpr1, mean_tpr1, 'k:',\n",
    "#         label='Mean ROC of $Caffe_6$ (AUC = %0.2f)' % mean_auc1, lw=3,c=palette[0])\n",
    "# plt.plot(mean_fpr, mean_tpr, 'k-.',\n",
    "#          label='Mean ROC of $Caffe_5$ (AUC = %0.2f)' % mean_auc, lw=3, c=palette[1])\n",
    "plt.plot(mean_fpr3, mean_tpr3, 'k-',\n",
    "         label='Mean ROC of fc6 layer (AUC = %0.2f$\\pm$%0.2f, acc=%0.2f$\\pm$%0.2f)' % (mean_auc3,std_AUC,mean_accuracy,std_accuracy), lw=3, c='black')\n",
    "\n",
    "plt.plot([0, 1], [0, 1], '--', color=(0.6, 0.6, 0.6), label='Luck')\n",
    "\n",
    "plt.xlim([-0.05, 1.05])\n",
    "plt.ylim([-0.05, 1.05])\n",
    "plt.xlabel('False Positive Rate',fontsize=20)\n",
    "plt.ylabel('True Positive Rate',fontsize=20)\n",
    "plt.title('ROC curves for fc6 layer',fontsize=20)\n",
    "plt.legend(loc=4,prop={'size':11})\n",
    "plt.grid('off')\n",
    "plt.savefig('/Users/Bato/Dropbox/tsne_python_5.25/Feature/CaffeNet/fine-tuning/OFF-the-shell/test1_3/fc6/fc6-AUC.png')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAGFCAYAAADTpKXkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x125a14490>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "layers = ['conv1','conv2','conv3','conv4','conv5','fc6','fc7','fc8']\n",
    "x = np.arange(len(layers))\n",
    "mean_auc = [0.56, 0.77, 0.80, 0.83, 0.91, 0.81, 0.90, 0.91]\n",
    "\n",
    "mpl.rc('xtick',labelsize=15)\n",
    "mpl.rc('figure',figsize = (6,6))\n",
    "plt.xticks(x, layers)\n",
    "plt.plot(x, mean_auc)\n",
    "plt.xlabel('Layers of CaffeNet')\n",
    "plt.ylabel('Mean AUC')\n",
    "plt.savefig('/Users/Bato/Dropbox/tsne_python_5.25/Feature/CaffeNet/all_mean_auc.png')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [default]",
   "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
}