721 lines (720 with data), 160.5 kB

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-30T10:17:23.037259Z",
     "start_time": "2017-11-30T10:17:20.465395Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "/Users/attia/Desktop/Work/workenv/lib/python3.6/importlib/_bootstrap.py:219: RuntimeWarning: compiletime version 3.5 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.6\n",
      "  return f(*args, **kwds)\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation, Flatten, Reshape\n",
    "from keras.layers import Conv2D, MaxPooling2D, AveragePooling2D\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras import regularizers\n",
    "from keras.losses import mean_squared_error\n",
    "import glob\n",
    "import matplotlib.patches as patches\n",
    "import json\n",
    "import numpy as np\n",
    "from matplotlib.path import Path\n",
    "import dicom\n",
    "import cv2\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The first step of the algorithm is to locate the Left-Ventricle and compute the Regions of Interest (ROI) around it. We want to re-create the convolutional neural network for this automatic detection. This CNN will generate binary mask with the ROI.  \n",
    "The first part is to open the dataset (pictures with the respective contour around the left ventricle) and generate the binary mask ROI (image with black background and a white foreground corresponding to the ROI) from the labeled data (contours drawn by experts), as it's the output of the CNN.\n",
    "The second part is the CNN model and the training."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Open dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-30T10:19:19.246479Z",
     "start_time": "2017-11-30T10:19:19.233486Z"
    }
   },
   "outputs": [],
   "source": [
    "def get_roi(image, contour, shape_out = 32):\n",
    "    \"\"\"\n",
    "    Create a binary mask with ROI from contour. \n",
    "    Extract the maximum square around the contour.\n",
    "    :param image: input image (needed for shape only)\n",
    "    :param contour: numpy array contour (d, 2)\n",
    "    :return: numpy array mask ROI (shape_out, shape_out)\n",
    "    \"\"\"\n",
    "    X_min, Y_min = contour[:,0].min(), contour[:,1].min()\n",
    "    X_max, Y_max = contour[:,0].max(), contour[:,1].max()  \n",
    "    w = X_max - X_min\n",
    "    h = Y_max - Y_min\n",
    "    mask_roi = np.zeros(image.shape)\n",
    "    if w > h :\n",
    "        mask_roi[int(Y_min - (w -h)/2):int(Y_max + (w -h)/2), int(X_min):int(X_max)] = 1.0\n",
    "    else :\n",
    "        mask_roi[int(Y_min):int(Y_max), int(X_min - (h-w)/2):int(X_max + (h -w)/2)] = 1.0\n",
    "    return cv2.resize(mask_roi, (shape_out, shape_out), interpolation = cv2.INTER_NEAREST)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-30T10:25:09.575124Z",
     "start_time": "2017-11-30T10:24:50.857652Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset shape : (495, 64, 64, 1) (495, 1, 32, 32)\n"
     ]
    }
   ],
   "source": [
    "X, X_fullsize, Y, contour_mask = create_dataset(n_set='train')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-30T10:23:31.873692Z",
     "start_time": "2017-11-30T10:23:31.746416Z"
    }
   },
   "outputs": [],
   "source": [
    "def create_dataset(image_shape=64, n_set='train', original_image_shape=256, \n",
    "                   roi_shape=32, data_path='./Data/'):\n",
    "    \"\"\"\n",
    "    Creating the dataset from the images and the contour for the CNN.\n",
    "    :param image_shape: image dataset desired size\n",
    "    :param original_image_shape: original image size\n",
    "    :param roi_shape: binary ROI mask shape\n",
    "    :param data_path: path for the dataset\n",
    "    :return: correct size image dataset, full size image dataset, label (contours) dataset\n",
    "    \"\"\"\n",
    "    \n",
    "    if n_set == 'train':\n",
    "        number_set = 3\n",
    "        name_set = 'Training'\n",
    "    elif n_set == 'test':\n",
    "        number_set = 1\n",
    "        name_set = 'Online'         \n",
    "    # Create dataset\n",
    "    series = json.load(open('series_case.json'))[n_set]\n",
    "    images, images_fullsize, contours, contour_mask = [], [], [], []\n",
    "    # Loop over the series\n",
    "    for case, serie in series.items():\n",
    "        image_path_base = data_path + 'challenge_%s/%s/IM-%s' % (name_set.lower(),case, serie)\n",
    "        contour_path_base = data_path + 'Sunnybrook Cardiac MR Database ContoursPart%s/\\\n",
    "%sDataContours/%s/contours-manual/IRCCI-expert/' % (number_set, name_set, case)\n",
    "        contours_list = glob.glob(contour_path_base + '*')\n",
    "        contours_list_series = [k.split('/')[7].split('-')[2] for k in contours_list]\n",
    "        # Loop over the contours/images\n",
    "        for c in contours_list_series:\n",
    "            # Get contours and images path\n",
    "            idx_contour = contours_list_series.index(c)\n",
    "            image_path = image_path_base + '-%s.dcm' % c\n",
    "            contour_path = contours_list[idx_contour]\n",
    "\n",
    "            # open image as numpy array and resize to (image_shape, image_shape)\n",
    "            image_part = dicom.read_file(image_path).pixel_array  \n",
    "\n",
    "            # open contours as numpy array\n",
    "            contour = []\n",
    "            file = open(contour_path, 'r') \n",
    "            for line in file: \n",
    "                contour.append(tuple(map(float, line.split())))\n",
    "            contour = np.array(contour)\n",
    "            # append binary ROI mask \n",
    "            contours.append(get_roi(image_part, contour))\n",
    "\n",
    "            # create mask contour with experts contours\n",
    "            x, y = np.meshgrid(np.arange(256), np.arange(256)) # make a canvas with coordinates\n",
    "            x, y = x.flatten(), y.flatten()\n",
    "            points = np.vstack((x,y)).T \n",
    "            p = Path(contour) # make a polygon\n",
    "            grid = p.contains_points(points)\n",
    "            mask_contour = grid.reshape(256,256)\n",
    "            mask_contour=mask_contour*1\n",
    "            contour_mask.append(mask_contour)\n",
    "            \n",
    "            # Open image and resize it \n",
    "            images.append(cv2.resize(image_part, (image_shape, image_shape)))\n",
    "            images_fullsize.append(cv2.resize(image_part, (original_image_shape, original_image_shape)))\n",
    "    X_fullsize = np.array(images_fullsize)\n",
    "    X = np.reshape(np.array(images), [len(images), image_shape, image_shape, 1])\n",
    "    Y = np.reshape(np.array(contours), [len(contours), 1, roi_shape, roi_shape])\n",
    "    print('Dataset shape :', X.shape, Y.shape)\n",
    "    return X, X_fullsize, Y, contour_mask"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For the input, the original image size of $256\\times256$ has been downsampled to $64\\times64$. The binary ROI mask (used as the labels) has a size of $32\\times32$ (as the CNN output mask). The white foreground of the binary mask is centered at the center of the left ventricle contour (known from the training manual contours)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "###### Note à modifier : \n",
    "Est ce qu'il faut shuffle (X,Y) avant le training ou le CNN le fait automatiquement ?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-22T14:01:10.120735Z",
     "start_time": "2017-11-22T14:00:51.350704Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset shape : (495, 64, 64, 1) (495, 1, 32, 32)\n"
     ]
    }
   ],
   "source": [
    "X, X_fullsize, Y, contour_mask = create_dataset()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-22T14:01:10.508510Z",
     "start_time": "2017-11-22T14:01:10.123522Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlMAAAEtCAYAAAAsgeXEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x123cfb668>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, ax = plt.subplots(ncols = 2, figsize=(10,10))\n",
    "ax[0].imshow(X[30].reshape(64,64))\n",
    "ax[0].set_title('Input image')\n",
    "ax[1].imshow(Y[30].reshape(32, 32))\n",
    "_ = ax[1].set_title('Input binary ROI mask')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-17T17:22:18.817523Z",
     "start_time": "2017-11-17T17:22:18.815186Z"
    }
   },
   "source": [
    "### Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-22T14:01:10.520421Z",
     "start_time": "2017-11-22T14:01:10.510002Z"
    }
   },
   "outputs": [],
   "source": [
    "def create_model(input_shape=(64, 64)):\n",
    "    \"\"\"\n",
    "    Simple convnet model : one convolution, one average pooling and one fully connected layer:\n",
    "    :return: Keras model\n",
    "    \"\"\"\n",
    "    model = Sequential()\n",
    "    model.add(Conv2D(100, (11,11), padding='valid', strides=(1, 1), input_shape=(input_shape[0], input_shape[1], 1)))\n",
    "    model.add(AveragePooling2D((6,6)))\n",
    "    model.add(Reshape([-1, 8100]))\n",
    "    model.add(Dense(1024, activation='sigmoid', kernel_regularizer=regularizers.l2(0.0001)))\n",
    "    model.add(Reshape([-1, 32, 32]))\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-22T14:01:10.699074Z",
     "start_time": "2017-11-22T14:01:10.524960Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Size for each layer :\n",
      "Layer, Input Size, Output Size\n",
      "Conv2D_1 (None, 64, 64, 1) (None, 54, 54, 100)\n",
      "Average_Pooling2D_1 (None, 54, 54, 100) (None, 9, 9, 100)\n",
      "Reshape_1 (None, 9, 9, 100) (None, 1, 8100)\n",
      "Dense_1 (None, 1, 8100) (None, 1, 1024)\n",
      "Reshape_2 (None, 1, 1024) (None, 1, 32, 32)\n"
     ]
    }
   ],
   "source": [
    "m = create_model()\n",
    "m.compile(loss='mean_squared_error',\n",
    "          optimizer='adam',\n",
    "          metrics=['accuracy'])\n",
    "print('Size for each layer :\\nLayer, Input Size, Output Size')\n",
    "for p in m.layers:\n",
    "    print(p.name.title(), p.input_shape, p.output_shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Training the CNN involves obtaining the optimum values of filters. In the paper, they are using a sparse autoencoder which acts a pre-training step because the number of training and labeled data is limited. The task of a autoencoder is to construct $x^{(i)}$ at the output from the hidden values. We didn't implement this sparse autoencoder. We can try data augmentation : the number of training and labeled data is limited so we can try to use data augmentation in order to generate new picture from our training set. With the Keras library, we can apply transformations on our training pictures to generate \"new\" pictures."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-22T14:01:16.175368Z",
     "start_time": "2017-11-22T14:01:16.163650Z"
    }
   },
   "outputs": [],
   "source": [
    "def training(m, X, Y, batch_size=16, epochs= 10, data_augm=False):\n",
    "    \"\"\"\n",
    "    Training CNN with the possibility to use data augmentation\n",
    "    :param m: Keras model\n",
    "    :param X: training pictures\n",
    "    :param Y: training binary ROI mask\n",
    "    :return: history\n",
    "    \"\"\"\n",
    "    if data_augm:\n",
    "        datagen = ImageDataGenerator(\n",
    "            featurewise_center=False,  # set input mean to 0 over the dataset\n",
    "            samplewise_center=False,  # set each sample mean to 0\n",
    "            featurewise_std_normalization=False,  # divide inputs by std of the dataset\n",
    "            samplewise_std_normalization=False,  # divide each input by its std\n",
    "            zca_whitening=False,  # apply ZCA whitening\n",
    "            rotation_range=50,  # randomly rotate images in the range (degrees, 0 to 180)\n",
    "            width_shift_range=0.1,  # randomly shift images horizontally (fraction of total width)\n",
    "            height_shift_range=0.1,  # randomly shift images vertically (fraction of total height)\n",
    "            horizontal_flip=True,  # randomly flip images\n",
    "            vertical_flip=False) \n",
    "        datagen.fit(X)\n",
    "        history = m.fit_generator(datagen.flow(X, Y,\n",
    "                                    batch_size=batch_size),\n",
    "                                    steps_per_epoch=X.shape[0] // batch_size,\n",
    "                                    epochs=epochs)         \n",
    "    else:\n",
    "        history = m.fit(X, Y, batch_size=batch_size, epochs=epochs)\n",
    "    return history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-22T14:03:04.978452Z",
     "start_time": "2017-11-22T14:01:20.441726Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "495/495 [==============================] - 5s 11ms/step - loss: 0.1845 - acc: 0.1955\n",
      "Epoch 2/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.1044 - acc: 0.1415\n",
      "Epoch 3/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0782 - acc: 0.0820\n",
      "Epoch 4/20\n",
      "495/495 [==============================] - 6s 11ms/step - loss: 0.0632 - acc: 0.0460\n",
      "Epoch 5/20\n",
      "495/495 [==============================] - 5s 11ms/step - loss: 0.0539 - acc: 0.0403\n",
      "Epoch 6/20\n",
      "495/495 [==============================] - 6s 11ms/step - loss: 0.0480 - acc: 0.0490\n",
      "Epoch 7/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0425 - acc: 0.0562\n",
      "Epoch 8/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0394 - acc: 0.0598\n",
      "Epoch 9/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0374 - acc: 0.0674\n",
      "Epoch 10/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0353 - acc: 0.0506\n",
      "Epoch 11/20\n",
      "495/495 [==============================] - 5s 11ms/step - loss: 0.0340 - acc: 0.0561\n",
      "Epoch 12/20\n",
      "495/495 [==============================] - 5s 11ms/step - loss: 0.0333 - acc: 0.0548\n",
      "Epoch 13/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0322 - acc: 0.0436\n",
      "Epoch 14/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0308 - acc: 0.0396\n",
      "Epoch 15/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0293 - acc: 0.0436\n",
      "Epoch 16/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0305 - acc: 0.0465\n",
      "Epoch 17/20\n",
      "495/495 [==============================] - 5s 11ms/step - loss: 0.0311 - acc: 0.0461\n",
      "Epoch 18/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0303 - acc: 0.0545\n",
      "Epoch 19/20\n",
      "495/495 [==============================] - 5s 10ms/step - loss: 0.0306 - acc: 0.0489\n",
      "Epoch 20/20\n",
      "495/495 [==============================] - 6s 12ms/step - loss: 0.0301 - acc: 0.0615\n"
     ]
    }
   ],
   "source": [
    "h = training(m, X, Y, batch_size=16, epochs= 20, data_augm=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-20T10:56:31.654864Z",
     "start_time": "2017-11-20T10:56:31.514178Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Learning curve')"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEWCAYAAABMoxE0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x12a43fbe0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "metric = 'loss'\n",
    "plt.plot(range(len(h.history[metric])), h.history[metric])\n",
    "plt.ylabel(metric)\n",
    "plt.xlabel('epochs')\n",
    "plt.title(\"Learning curve\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The original MR image size is $256\\times256$. Therefore, the output mask is up-sampled from $32\\times32$ to this original MR size. The center of the mask is then computed and used to crop a ROI of size $100\\times100$ from the original image for further processing in the next stage (stacked AE)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Predictions and computations for next stage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-20T10:56:37.971243Z",
     "start_time": "2017-11-20T10:56:36.936688Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Stored 'y_pred' (ndarray)\n"
     ]
    }
   ],
   "source": [
    "y_pred = m.predict(X, batch_size=16)\n",
    "%store y_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-20T10:56:39.666378Z",
     "start_time": "2017-11-20T10:56:39.640956Z"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def compute_roi_pred(y_pred, idx, roi_shape=32):\n",
    "    \"\"\"\n",
    "    Computing and cropping a ROI from the original image for further processing in the next stage\n",
    "    :param y_pred: predictions\n",
    "    :param idx: desired image prediction index\n",
    "    :param roi_shape: shape of the binary mask\n",
    "    \"\"\"\n",
    "    # up sampling from 32x32 to original MR size\n",
    "    pred = cv2.resize(y_pred[idx].reshape((roi_shape, roi_shape)), (\n",
    "                      256,256), cv2.INTER_NEAREST)\n",
    "    # select the non null pixels\n",
    "    pos_pred = np.array(np.where(pred > 0.5))\n",
    "    # get the center of the mask\n",
    "    X_min, Y_min = pos_pred[0, :].min(), pos_pred[1, :].min()\n",
    "    X_max, Y_max = pos_pred[0, :].max(), pos_pred[1, :].max()  \n",
    "    X_middle = X_min + (X_max - X_min) / 2\n",
    "    Y_middle = Y_min + (Y_max - Y_min) / 2\n",
    "    # Find ROI coordinates\n",
    "    X_top = int(X_middle - 50)\n",
    "    Y_top = int(Y_middle - 50)\n",
    "    X_down = int(X_middle + 50)\n",
    "    Y_down = int(Y_middle + 50)\n",
    "    # crop ROI of size 100x100\n",
    "    mask_roi = np.zeros((256, 256))\n",
    "    mask_roi = cv2.rectangle(mask_roi, (X_top, Y_top), (X_down, Y_down), 1, -1)*255\n",
    "    return X_fullsize[idx][X_top:X_down, Y_top:Y_down], mask_roi, contour_mask[idx][X_top:X_down, Y_top:Y_down]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-20T10:56:40.603766Z",
     "start_time": "2017-11-20T10:56:40.600143Z"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pred2, mask_roi, mask_contour = compute_roi_pred(y_pred, 234)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-20T10:56:41.795082Z",
     "start_time": "2017-11-20T10:56:41.457358Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x116d6e0b8>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlkAAADHCAYAAAAqC0ZdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10bc7ad30>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f, ax = plt.subplots(ncols=3, figsize=(10,10))\n",
    "ax[0].imshow(X_fullsize[234])\n",
    "ax[1].imshow(mask_roi)\n",
    "ax[2].imshow(pred2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-11-20T10:57:50.028865Z",
     "start_time": "2017-11-20T10:57:50.023748Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cv2.imwrite('./Rapport/images/mask_roi.png', mask_roi)\n",
    "cv2.imwrite('./Rapport/images/predic_roi.png', pred2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(<matplotlib.image.AxesImage at 0x1251b86a0>, (100, 100))"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nLzXI4EsNMvhSgwy+1CCDLzXI4EsN+j+YdxuADh3nuAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x124fee6d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(mask_contour), mask_contour.shape"
   ]
  },
  {
   "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": "3.6.3"
  },
  "toc": {
   "nav_menu": {
    "height": "84px",
    "width": "252px"
   },
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": "block",
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}