pytorch_loaders.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "56f95dfd",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "from torchvision.io import read_image\n",
    "# for reading and displaying images\n",
    "from skimage.io import imread\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# for creating validation set\n",
    "from sklearn.model_selection import train_test_split\n",
    "# for evaluating the model\n",
    "from sklearn.metrics import accuracy_score\n",
    "import nibabel as nib\n",
    "import glob\n",
    "import random\n",
    "import numpy as np\n",
    "import os\n",
    "from natsort import natsorted\n",
    "\n",
    "# PyTorch libraries and modules\n",
    "import torch\n",
    "from torch.autograd import Variable\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "from torch.optim import *\n",
    "import h5py\n",
    "from torch.utils.data import Dataset\n",
    "from torchvision import datasets\n",
    "from torchvision.transforms import ToTensor\n",
    "import matplotlib.pyplot as plt\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "a5f32c8d",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "training_data = datasets.FashionMNIST(\n",
    "    root=\"data\",\n",
    "    train=True,\n",
    "    download=True,\n",
    "    transform=ToTensor()\n",
    ")\n",
    "\n",
    "test_data = datasets.FashionMNIST(\n",
    "    root=\"data\",\n",
    "    train=False,\n",
    "    download=True,\n",
    "    transform=ToTensor()\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "ffd7e8eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "#Reading an example PCT volume\n",
    "LOCATION = 'HOME'\n",
    "\n",
    "if LOCATION == 'HOME':\n",
    "    ROOT_EXP = '/Users/admin/work/aifnet_replication-main'\n",
    "    root_dir  = '/Users/admin/work/data/'\n",
    "\n",
    "\n",
    "if LOCATION == 'LOCAL':\n",
    "    ROOT_EXP = '/Users/sebastianotalora/work/postdoc/ctp/aifnet_replication/'\n",
    "    root_dir  = '/Users/sebastianotalora/work/postdoc/data/ISLES/'\n",
    "\n",
    "if LOCATION == 'INSEL':\n",
    "    ROOT_EXP = '/home/sebastian/experiments/aifnet_replication/'\n",
    "    root_dir  = '/media/sebastian/data/ASAP/ISLES2018_Training'\n",
    "\n",
    "if LOCATION == 'SERVER':\n",
    "    ROOT_EXP = '/home/sotalora/aifnet_replication/'\n",
    "    root_dir     = '/data/images/sotalora/ISLES18/'\n",
    "\n",
    "aif_annotations_path = ROOT_EXP + 'radiologist_annotations.csv'\n",
    "\n",
    "\n",
    "aif_annotations_path = ROOT_EXP + '/radiologist_annotations_cleaned.csv'#'radiologist_annotations.csv'#'annotated_aif_vof_complete_revised.csv'\n",
    "min_num_volumes_ctp = 43\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "d5684881",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "class ISLESDataset(Dataset):\n",
    "    def __init__(self, annotations_file, partition_file_path, \n",
    "                 minimum_number_volumes_ctp, img_dir, vascular_fn='AIF',\n",
    "                 transform=None, target_transform=None):\n",
    "        self.img_dir = img_dir\n",
    "        self.transform = transform\n",
    "        self.target_transform = target_transform\n",
    "        self.minimum_number_volumes_ctp = minimum_number_volumes_ctp\n",
    "        dataset_dir = os.path.join(img_dir, \"TRAINING\")\n",
    "        dataset_dir_test = os.path.join(img_dir, \"TESTING\")\n",
    "        self.filenames_4D = natsorted(glob.glob(dataset_dir + \"/case_*/*4D*/*nii*\") + glob.glob(dataset_dir_test + \"/case_*/*4D*/*nii*\"))\n",
    "        self.cases_paths = {path.split('.')[-2]: path for path in self.filenames_4D}\n",
    "        self.vascular_fn = vascular_fn\n",
    "        partition_file = open(partition_file_path,'r')\n",
    "        relevant_cases = []\n",
    "        for line in partition_file:\n",
    "            relevant_cases.append(line.split('.')[-2])\n",
    "        partition_file.close()\n",
    "        self.fold_img_paths = {key: self.cases_paths[key] for key in relevant_cases}\n",
    "        self.fold_ids = [int(valid_indx) for valid_indx in list(self.fold_img_paths.keys())]\n",
    "        self.all_labels = pd.read_csv(aif_annotations_path)\n",
    "        self.img_labels = self.all_labels[self.all_labels['imageID'].isin(self.fold_ids)]\n",
    "        #print(self.fold_img_paths)\n",
    "        \n",
    "    def __len__(self):\n",
    "        return len(self.fold_img_paths)\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        img_id = str(self.img_labels.iloc[idx]['imageID'])\n",
    "        img_path = self.fold_img_paths[img_id]\n",
    "        image = read_volume_sequence(img_path,self.minimum_number_volumes_ctp)\n",
    "        AIF = image[self.img_labels.iloc[idx]['AIFx'],self.img_labels.iloc[idx]['AIFy'],self.img_labels.iloc[idx]['AIFz'],:]\n",
    "        VOF = image[self.img_labels.iloc[idx]['VOFx'],self.img_labels.iloc[idx]['VOFy'],self.img_labels.iloc[idx]['VOFz'],:]\n",
    "        if self.transform:\n",
    "            image = self.transform(image)\n",
    "        if self.target_transform:\n",
    "            label = self.target_transform(label)\n",
    "        label = AIF #By default return only the AIF\n",
    "        if self.vascular_fn == 'VOF':\n",
    "            label = VOF\n",
    "        return image, label\n",
    "def read_volume_sequence(fname,minimum_number_volumes_ctp):\n",
    "        cur_nib = nib.load(fname)\n",
    "        ctp_vals = cur_nib.get_fdata()\n",
    "        volume_sequence = ctp_vals[:,:,:,0:minimum_number_volumes_ctp]\n",
    "        return volume_sequence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "674ceed8",
   "metadata": {},
   "outputs": [],
   "source": [
    "annotations_file = aif_annotations_path\n",
    "current_fold = 1\n",
    "minimum_number_volumes_ctp = 43\n",
    "partition_file_path = ROOT_EXP + '/partitions_cleaned/fold_'+str(current_fold)+'/train_v2.txt'\n",
    "training_data = ISLESDataset(\n",
    "annotations_file, partition_file_path, \n",
    "                 minimum_number_volumes_ctp, root_dir, vascular_fn='AIF',\n",
    "                 transform=None, target_transform=None\n",
    ")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "67076ec6",
   "metadata": {},
   "outputs": [],
   "source": [
    "sample_idx = torch.randint(10, size=(1,)).item()\n",
    "img, label = training_data[sample_idx]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "d059c772",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(43,)"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array(label).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "550485fa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Normalized Density (HU)')"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd3ic5ZXw4d9R771YxXKRe2+ADRiMicGAY4gDrCHUhJBsEtgUEiDJbthN3c2SXgj7xUACGAi9BTDFVINx77Zsy5YlyyqW1fvofH/MSJZtaTSSZjSS5tzXNZc077wz79HYmqPnOU8RVcUYY4wBCPJ3AMYYYwYPSwrGGGM6WFIwxhjTwZKCMcaYDpYUjDHGdLCkYIwxpoMlBWOMMR0sKRjTTyKyVkROiEh4p2MPi8hPXN+PFhEVkdpOt63+i9iY7llSMKYfRGQ0sBBQYHkPpyeoaozrNtPXsRnTF5YUjOmfm4CPgYeBm/0bijH9F+LvAIwZ4m4CfgV8AnwsIumqWuLnmIzpM2spGNNHInI+MAp4SlU3AgeA6908pVxEKl23uwYkSGN6yVoKxvTdzcAbqlruuv+469ivuzk/RVVbByQyY/rIkoIxfSAikcC1QLCIHHMdDgcSRMSKyGbIsqRgTN9cBTiA6UBzp+NP4awzGDMkWU3BmL65GXhIVQtU9Vj7DfgD8AXsDy4zRIltsmOMMaadtRSMMcZ0sKRgjDGmgyUFY4wxHSwpGGOM6TCkR0ikpKTo6NGj/R2GMcYMKRs3bixX1dSuHhvSSWH06NFs2LDB32EYY8yQIiKHu3vMuo+MMcZ0sKRgjDGmgyUFY4wxHYZ0TcEYY3yhpaWFwsJCGhsb/R1Kv0RERJCdnU1oaKjHz7GkYIwxpyksLCQ2NpbRo0cjIv4Op09UlePHj1NYWMiYMWM8fp51HxljzGkaGxtJTk4esgkBQERITk7udWvHkoIxxnRhKCeEdn35Gaz7yPRa4Yl6XtpaTEZ8BCOTIhmZGEVqbPiw+CUyJtBZUjC99vNX9/DK9uJTjoWHBJGdGElOUhTj0mL4+kXjSIgK81OExgwPzz33HCtWrGD37t1MmjSJQ4cOsWzZMnbs2MHatWu58sorO+oFKSkpvPnmm/2+piUF0ysl1Y28vvMYt5w7mhvm53CkooGCinqOVNRz5EQ9RyoaeD+vnPWHTvDYbecQE27/xYzpq9WrV3P++efzxBNPcN99953x+MKFC3n55Ze9ek37jTW9snp9Aa1tyi3njmZ0SjTj0mLPOOfNXSV85dGNfOnhT3n41rOJDAv2Q6TGDG21tbV8+OGHvPPOOyxfvrzLpOALlhSMx1ocbaxeX8CFE1IZnRLd7XmfmZLOr66dyTef3MK/PraRB2+cR1iIjWkwQ9N/vrSTXUervfqaUzLj+NFnp7o95/nnn2fp0qVMmDCBpKQkNm3aRFJS0innvP/++8yaNQuAa665hh/84Af9js1+U43H1uwqoaS6iZsWjOrx3CtnZfHzz01n7d4yvvnkZlodbQMQoTHDx+rVq1m5ciUAK1euZPXq1Wecs3DhQrZs2cKWLVu8khDAWgqmF/627hBZCZEsmpjm0fkrz86htqmVn7yym8jQ7fzy6hkEBdkIJTO09PQXvS8cP36ct99+mx07diAiOBwORISvfe1rPr+2tRSMR/JKavj4YAU3zB9FcC8+2G9bOJZvfWYCz2wq5L6XdqKqPozSmOHh6aef5qabbuLw4cMcOnSII0eOMGbMGAoLC31+bWspGI/8/ePDhAUHce287F4/986Lx1Hb1ML/vZ9PdHgIdy+d5IMIjRk+Vq9ezT333HPKsc9//vP87Gc/8/m1LSmYHtU2tfLspiKWzcggOSa8188XEb5/+WRqmxz8ee0BxqfFsGJO75OLMYFi7dq1Zxy78847ufPOOzvuL1q0iEWLFnn92tZ9ZHr03OYiaptaucGDAnN3RISfXDWNnKQoXj1t4psxZvCwpGDcUlUeXXeYaVlxzB6Z0K/XCg4Szs1N5pP8ChxtVlswZjCypGDcWp9fwd6SGm6cP8oraxstyE2mprGVnUervBCdMb4zHAZF9OVnsKRg3Prbx4eJjwxl+cwsr7zegrHJAKw7cNwrr2eML0RERHD8+PEhnRja91OIiIjo1fOs0Gy6VVrdyOs7nOsceWupirS4CHJTo1l38DhfuTDXK69pjLdlZ2dTWFhIWVmZv0Ppl/ad13rDkoLp1ur1R2htU74wv+8F5q4syE3muU1FtDjaCA22xqoZfEJDQ3u1W9lwYr+RpkstjjYeX3+YCyakMsbNOkd9sWBsCnXNDrYXWV3BmMHGkoLp0puudY5u9HIrAWD+WOeiXlZXMGbwsaRgzrD3WA0//+ceshIiWTzJs3WOeiM5JpyJ6bGWFIwZhCwpmFM8t7mQq/74IQ0tDn533axerXPUGwtyk9lwuIKmVodPXt8Y0zc+SwoiskpESkVkRxeP3SUiKiIpnY7dKyL7RWSviFzqq7hM15paHfzw+e1868mtTM+O55U7zmfuqKSen9hHC3KTaWxpY+sRqysYM5j4sqXwMLD09IMiMhJYAhR0OjYFWAlMdT3nTyJi23UNkMIT9Vz7wDoe/biAr1wwlsdvO4e0uN6Nbe6t+WOSEbG6gjGDjc+Sgqq+B1R08dCvge8BnWeFXAk8oapNqpoP7AfO9lVs5qS1e0tZ9vsPOFhWxwM3zOXeyycTMgDDROOjQpmSEce6g+U+v5YxxnMDWlMQkeVAkapuPe2hLOBIp/uFrmNdvcbtIrJBRDYM9Ykl/vantfu59eFPGREXwYt3nM/SaSMG9PoLxiazqaCSxharKxgzWAxYUhCRKOAHwH909XAXx7qcX66qD6rqPFWdl5qa6s0QA0pdUyu/fH0vF09K57mvnef1uQieWJCbTHNrG5sKTgz4tY0xXRvIlkIuMAbYKiKHgGxgk4iMwNkyGNnp3Gzg6ADGFnDySmtRhWvmZXttCYveOmtMEkFWVzBmUBmwpKCq21U1TVVHq+ponIlgjqoeA14EVopIuIiMAcYD6wcqtkC0r6QGgAnpsX6LIS4ilOlZ8ZYUjBlEfDkkdTWwDpgoIoUi8qXuzlXVncBTwC7gNeDrqmodzT6UV1JDeEgQOUlRfo1jfm4yWwsrqW9u9WscxhgnX44+uk5VM1Q1VFWzVfWvpz0+WlXLO93/qarmqupEVf2nr+IyTvtKaslNjfHZ5DRPLRibTItD2XDI6grGDAY2ozlA5ZXUMCE9xt9hcNboJEKChHUHrQvJmMHAkkIAqmls4WhVI+P9WE9oFx0ewsyRCVZXMGaQsKQQgPaX1gIwPs3/LQVwdiFtL6qiprHF36EYE/AsKQSgvBJnUvDnyKPOFuQm42hTPj3U1QR4Y8xAsqQQgPa5Rh6N9PPIo3ZzRyUSFhxkXUjGDAKWFALQvtJaxqX5f+RRu4jQYGblJFix2ZhBwJJCAHKOPBocXUftFoxNZufRaqrqra5gjD9ZUggw1Y0tFFc1Mn4QDEftbEFuMqrwSb61FozxJ0sKAaajyJw2uFoKs3MSCA8J4iOrKxjjV5YUAsz+UueaR4OtpRAeEszC8amsXl/AB3m2x4Ix/mJJIcDsK6klIjSIkYmDY+RRZ/9z9QzGpETzpUc+tcRgjJ9YUggw+0pqGJcWQ9AgGXnUWVJ0GI9/eX5HYvhwvyUGYwaa26QgItkicpeIvCAin4rIeyLyJxG5QkQsoQxBeSW1g66e0FlSdBiP3XYOo5OdieEjSwzGDKhuP9hF5CFgFdAM/DdwHfA14E1gKfCBiFwwEEEa76hqaOFY9eBY88id5JhwHvvyOeQkRfHFRz61SW3GDCB3f+3fr6qXqOrvVPUjVd2vqjtU9VlVvQNYhO2ONqS0F5kHw+qoPUmJCefxL89nZGIUX3z4Uz62iW3GDIhuk4Kq7nD3RFVtVtX93g/J+Mq+QbbmUU/aE0N2YiS3PmSJwZiB4K77aLuIbOt02yoib4nID0UkYiCDNN6xr6SGyNBgshIi/R2Kx1JjnYkhKzGS2/+2gRZHm79DMmZYc9d9tAz4bKfbcuAuIAX4ve9DM96237Xm0WAceeROamw4dyweR3Vja8ey38YY3wjp7gFVPdzF4cPAZhHZ7LuQjK/sK6nhvHEp/g6jT6ZmxgOwo6iKyRlxfo7GmOGrr8NKbTjqEFPV0EJJddOQqSecbkxKNFFhwew8Wu3vUIwZ1rptKYjInC4OJwI3AO/5LCLjE3klQ2fkUVeCg4QpGXHsPFrl71CMGda6TQrA/afdV+A4sBZ40FcBGd9oH3k0fhBPXOvJ1Mw4/rGxkLY2HXJ1EWOGCnc1hYsGMhDjW/tKaogKG1ojj043NSueR9YdJv94HbmpQ7PFY8xg56776Nvunqiqv/J+OMZX8kprGD8ERx51Ns1VbN55tNqSgjE+4q5gHNvpdtdp94duH0SA2ldSy7gh3HUEzuW+w4KD2FlkdQVjfMVd99F/tn8vIld1vm+Glsr6ZspqmoZskbldaHAQE0fEssOKzcb4jKdDS9WnURifyisdWstbuDMtK44dRdWo2n9JY3zBZ/MNRGSViJSKyI5Ox34pIntcy2Y8JyIJnR67V0T2i8heEbnUV3EFon0lg3O3tb6YmhlPVUMLhSca/B2KMcOSR2sfAZM6rYG03XWsJw/jXGK7szXANFWdAewD7nVdawqwEpjqes6fRCS49z+O6UpeSS3RQ3zkUbtpWSeLzcYY73M3T2FZf15YVd8TkdGnHXuj092Pgatd318JPKGqTUC+iOwHzgbW9ScG47SvpIZx6bGIDN2RR+0mjYglOEjYebSKpdNG+DscY4Ydd0mhQHvouBUR6ekcN74IPOn6PgtnkmhX6DrW1TVvB24HyMnJ6eOlA8u+kloumpjq7zC8IiI0mHGpMeywEUjG+IS7msI7InKHiJzyySsiYSKyWEQeAW7uy0VF5AdAK/BY+6EuTusy2ajqg6o6T1XnpaYOjw86XzpR10x57dBd86grU7Pi2GHdR8b4hLuksBRwAKtF5KiI7BKRg0Aezq05f62qD/f2giJyM86uqS90amUUAiM7nZaN7ermFe1F5nHDoMjcblpmPGU1TZRWN/o7FGOGHXfzFBqBP+Es+obi3EehQVUr+3oxEVkK3A1cqKr1nR56EXhcRH4FZALjgfV9vY45aTgNR23XudicFmf7PRnjTR4NSVXVFlUt7k1CEJHVOAvFE0WkUES+BPwB52zoNSKyRUQecL3+TuApYBfwGvB1VXX08mcxXcgrqSEmPITM+OHz4Tk5w5ngrK5gjPe5KzT3i6pe18Xhv7o5/6fAT30VT6ByLm8RMyxGHrWLjQhlTEq0zWw2xgdss5xhLq+0Zsgvb9GVqZlxNlfBGB/oMSmIyDdEJHEggjHeVVHXTHlt87CqJ7SblhVP4YkGKuub/R2KMcOKJy2FEcCnIvKUiCyV4dQPMYxVN7bw/We3Ayf3Nx5OpmY692m21oIx3tVjUlDVH+IcDfRX4BYgT0R+JiK5Po7N9NH2wiqW/e4D1uwu4fuXT2L+2CR/h+R17YnOis3GeJdHhWZVVRE5BhzDOeksEXhaRNao6vd8GaDxnKry948P85OXd5McE8ZTX5nP3FHDLyEAJEWHkZUQaS0FY7ysx6QgInfinLlcDvw/4Luq2iIiQTgnsllSGASqG1u495ntvLK9mIsmpvKra2eRGB3m77B8ampmnI1AMsbLPGkppAArVPVw54Oq2iYi/Vo0z3jHjqIqvv74JgpPNHDPZZO4feHYIb3tpqemZsazZncJtU2txIT7bHS1MQHFk0LzmNMTgoj8HUBVd/skKuOxdQeOs+LPH9HU0sYTt8/nqxfmBkRCAOeGO6qwu9i6kIzxFk+SwtTOd1z7HMz1TTimt17ZfpSw4CBeufN8zho9POsH3elY7sKKzcZ4jbtNdu4VkRpghohUu241QCnwwoBFaNzKK6llQnoMyTHh/g5lwKXFhpMSE24rphrjRd0mBVX9uarGAr9U1TjXLVZVk1X13gGM0biRV1o7LCeneUJEnMVmaykY4zXdVudEZJKq7gH+ISJzTn9cVTf5NDLTo/LaJirqmhkfoEkBnHWFD/aX09jiICLUdnA1pr/cDdn4Ns4dzu7v4jEFFvskIuOx9r0ShuPaRp6alhmPo03ZV1LDjOwEf4djzJDnbj+F211fLxq4cExv5JUMv70Sequ92LyjqNqSgjFe4MmCeNeISKzr+x+KyLMiMtv3oZme7CupITYihLTYwCsyt8tOjCQuIsQmsRnjJZ4MSf13Va0RkfOBS4FHgAd8G5bxRHuROZDXKHQWm+NtWKoxXuJJUmjfAe0K4M+q+gIwvNdPGAJUlbyS4blXQm9Ny4pj97EaWhxt/g7FmCHPk6RQJCJ/Aa4FXhWRcA+fZ3yovLaZE/UtjE8L3HpCu2lZ8TS3ttnMZmO8wJMP92uB14Glrj2ak4Dv+jQq06O8jpFHlhQWjk8lJjyE376Z5+9QjBnyPNlPoR7nDOY6EckBQoE9vg7MuGfDUU9Kig7jG4vH8daeUt7PK/N3OMYMaZ6MProDKAHWAK+4bi/7OC7Tg32ltcRFhJAawCOPOrv1vNGMTIrkJy/vptVqC8b0mSfdR/8GTFTVqao63XWb4evAjHvOInNgjzzqLDwkmO9fNpm9JTU8ueGIv8MxZsjyJCkcAWy83yCiquwrqQ3o5S26snTaCM4ek8Sv3thHdWOLv8MxZkjyJCkcBNa6Vk39dvvN14GZ7pXVNlHV0GL1hNOICP+xbAoV9c388e39/g7HmCHJk6RQgLOeEAbEdroZP7HlLbo3LSueq+dks+rDfA4fr/N3OMYMOT3uYaiq/wkgItGqar9lg0D7yKPxadZS6Mp3L53IK9uL+fmre3jgRtsPypje8GT00QIR2QXsdt2fKSJ/8nlkplv7SmqJjwy1kUfdSIuL4GuLcnlt5zHWHTju73CMGVI86T76Dc41j44DqOpW4IKeniQiq0SkVER2dDqWJCJrRCTP9TWx02P3ish+EdkrIpf2/kcJHO3LW9jIo+7dtnAsWQmR/PjlXTja1N/hGDNkeLRchaqePsbP0eWJp3oYWHrasXuAt1R1PPCW6z4iMgVYiXM/6KXAn1x7QZvTqCp5pTbyqCcRocHcfdkkdhVX88zGQn+HY8yQ4dGQVBE5F1ARCRORu3B1Jbmjqu8BFacdvhLnKqu4vl7V6fgTqtqkqvnAfuBsT36AQFNW4xp5ZPWEHn12RgZzchL4n9f3UtvU6u9wjBkSPEkKXwW+DmQBhcAs1/2+SFfVYgDX1zTX8Syc8yHaFbqOnUFEbheRDSKyoaws8JY02GcjjzwmIvz7simU1zbxyEeH/B2OMUOCJ2sflavqF1Q1XVXTVPUGVfV29a6rzvEuO4JV9UFVnaeq81JTU70cxuDXPvJonM1R8MjsnETmj03i6Y2FqFptwZieuE0KInKRa6e1na7b0yKyqB/XKxGRDNdrZwClruOFwMhO52UDR/txnWErr7SGhKhQUmNs5JGnVszJJr+8js1HKv0dijGDXrdJQUSuAFYBLwHXA18AXgVWicjlfbzei8DNru9vxrn6avvxlSISLiJjgPHA+j5eY1jbV1LLhDRb86g3Lps2gojQIJ7dZAVnY3rirqXwXeAqVX1IVbeq6hZVXYWzOHx3Ty8sIquBdcBEESkUkS8BvwCWiEgesMR1H1XdCTwF7AJeA76uqp6McAoo7butjbeuo16JjQjl0qkjeGlrMU2t9t/KGHfczWge4ZqTcApV3SYi6T29sKpe181DF3dz/k+Bn/b0uoGstKaJ6sZWKzL3wYo52byw5Sjv7Cll6bQMf4djzKDlrqXgbkkLW+7CDzqWt7CWQq+dl5tMWmw4z2wq8ncoxgxq7loKuSLyYhfHBRjro3iMG+3DUW1f5t4LCQ7iqtlZrPogn4q6ZpKiw/wdkjGDkrukcKWbx/7X24GYnuWV1JAYFUpKjH2g9cWKOVk8+N5BXtp6lJvPHe3vcIwZlLpNCqr67kAGYnq2r6SG8bbbWp9NGhHHlIw4nt1UaEnBmG54tPaR8T/nyKNa21inn1bMyWJrYRX7S2v8HYoxg5IlhSGipLqJmiYbedRfy2dlEhwkPGsFZ2O6ZElhiDi5sY4lhf5Ii43ggvEpPLe5iDZbUtuYM3RbUxCRl+hm/SEAVV3uk4hMl2w4qvesmJPNHas38/HB45w7LsXf4RgzqLhrKfwvcD+QDzQA/+e61QI73DzP+EBeSS1J0WGk2JpH/bZkSjqx4SE2Z8GYLnSbFFT1XdcIpNmq+i+q+pLrdj1w/sCFaAD2ldbYnsxeEhEazBUzMvjnjmLqm22fBWM686SmkCoiHZPVXAvWBd6a1X6kquwvqbUisxetmJNNfbOD13ce83coxgwqniSFbwFrRWStiKwF3gG+6dOozCmOVTe6Rh5ZS8Fb5o1KZGRSpI1CMuY07mY0A6Cqr4nIeGCS69AeVW3ybVims/blLcbZyCOvCQoSPjc7m9+/ncexqkZGxEf4OyRjBoUek4KIRAHfBkap6pdFZLyITFTVl30fngHn8haAtRS8bMXsLH73Vh6/XrOPiyalERkWTFRYMJGhwUS6viZFhxERGuzvUI0ZMD0mBeAhYCOwwHW/EPgHYElhgOwrqSE5OoxkG3nkVaNTolkwNpknNxzhyQ1HujwnLTacD+5eTFiITekxgcGTpJCrqv8iItcBqGqD2OI7A2p/aS25NvLIJx7+4lkcrWykodlBQ0srDc1t1De30tDiYHdxDQ+8e4BP8o+zcLyNrTCBwZOk0CwikbgmsolILmA1hQF06Hg9l0zpcV8j0wfhIcGMSYnu8rFLpzp4+KN81uwqsaRgAoYnbeL7cG6ROVJEHgPeAr7ny6DMSVUNLVTUNXf7wWV8JyI0mIXjU3lzVwmqtiSGCQw9JgVVfQNYAdwCrAbmqepa34Zl2h0qd25yN9qSgl8smZLO0apGdh6t9ncoxgyIHpOCiLwFnKOqr6jqy6paLiIPDkBsBjh03JkUrKXgHxdPSkME1uwq8XcoxgwIT7qPxgB3i8iPOh2b56N4zGnyy+sQgZykKH+HEpCSY8KZm5NoScEEDE+SQiVwMZAuIi+JSLyPYzKd5JfXkRkfaWPl/WjJlHR2FVdTVNng71CM8TlPkoKoaquqfg14BvgASPNtWKbdofI6RqdYK8GflrhGfr1prQUTADxJCg+0f6OqD+MsOL/ho3hMJ6pKfnkdo5OtnuBPY1NjGJsazZu7LSmY4a/bpCAica5v/yEiSe03nPsr3DUg0QW4E/UtVDe2WpF5EFgyJZ2PDx6nurHF36EY41PuWgqPu75uBDa4vm7sdN/4WH77cFRrKfjdksnptDiUtXvL/B2KMT7V7YxmVV3m+jpm4MIxnbXPURiTaknB32bnJJIcHcabu0pYPjPT3+EY4zPu9mie4+6JqrrJ++GYzg4dryNIYGSiFZr9LThIuHhyGv/ccYwWRxuhwbZAnhme3K19dL+bxxRY3NeLisi3gNtcr7MduBWIAp4ERgOHgGtV9URfrzEc5JfXkZ0YZSt0DhKfmZzOUxsKWZ9fwXnjUvwdjjE+4a776CJfXFBEsoA7gSmuFVefAlYCU4C3VPUXInIPcA9wty9iGCryy+tseYtBZOH4VMJDglizq8SSghm2PPoTVESmici1InJT+62f1w0BIkUkBGcL4ShwJfCI6/FHgKv6eY0hTVU5VF7HmGTrOhosIsOCWTg+hTW2QJ4ZxjxZ++hHwO9dt4uA/wGW9/WCqloE/C9QABQDVa5F99JVtdh1TjHdTJATkdtFZIOIbCgrG74jQcpqm6hrdlhLYZBZMiWdosoGdhfX+DsUY3zCk5bC1TiXuTimqrcCM4E+bwEmIok4WwVjgEwgWkRu8PT5qvqgqs5T1XmpqcN3jftD5fWArY462CyelG4L5JlhzZOk0KCqbUCra0JbKTC2H9f8DJCvqmWq2gI8C5wLlIhIBoDra2k/rjHkdQxHtTkKg0pqbDizRyawZvcxf4dijE94khQ2iEgC8H84J65tAtb345oFwHwRiXJt63kxsBt4EbjZdc7NwAv9uMaQl3+8jpAgITsx0t+hmNMsmTKCHUXVFFfZAnlm+PFkk52vqWqlqj4ALAFudnUj9YmqfgI8jTO5bHfF8CDwC2CJiOS5rvOLvl5jODhUXkdOUhQhNh5+0LEF8sxw5skezYjIDJzzB0Jc98ep6rN9vaiq/gj40WmHm3C2Ggw2HHUwy02NZkxKNGt2l3LjgtH+DscYr+oxKYjIKmAGsBNocx1WnLUA4wNtbcrh4/Wcm2tj4QcjEWHJlHQe+jCfn726m9qmVupcN+f3DhpbHFwzL5vbzh9LUJD4O2QzxDjalN+/ncfVc7PJHuAVDTxpKcxX1Sk+j8R0KKlppKHFwRjbR2HQWj4zk4c/OsTf1h0iJjyE6PAQosNCiAkPISUmjNqmVn726h4+OVjB/dfOJCEqzN8hmyFk4+ET/ObNPNoUvr1kwoBe25OksE5EpqjqLp9HY4BOq6Na99GgNS0rnr0/XopzrMSZVJW/f3yYH7+8iyt+9wF/uH42s3MSBzhKM1S9vcc5+HJbYeWAX9uTKuYjOBPDXhHZJiLbRWSbrwMLZB1zFGw46qDWXUJof+ymBaN5+qvnIgLX/mUdqz7It5nQxiPvuJLC9sKqAf8/40lSWAXcCCwFPgssc301PnLoeB1hwUFkJthw1KFu5sgEXrljIRdOSOO/Xt7Fvz66yTbqMW4VVTawt6SGsSnRHK9rHvC9wT1JCgWq+qKq5qvq4fabzyMLYPnldYxKjiLYCpTDQnxUKP9301x+cPlk3txdwrLffcCuo9X+DssMUu1dR99YPA5wthYGkidJYY+IPC4i14nIivabzyMLYIdsOOqwIyJ8+YKxPPmV+TS1OvjG45twtFlX0nD3wpYifvj89l495509peQkRXHFjAxCg4WtgzApROKcQ3AJzm6j9i4k4wNtbcrhinrbl3mYmjsqiX9fNoWD5XWs2TX4l8rYdbSai+9fS0l1o79DGaLhJlgAAB0ZSURBVHKqG1v40Ys7efTjAo9bhg3NDj7cX87iSWmEhwQzaUQc24sGttjsNimISDBQrqq3nnb74gDFF3COVjXQ3NpmReZh7LJpGYxKjuLP7x70WhHR0aZ8+6ktPLe50Cuv1+6JTws4UFbHxwePe/V1A8Ff3j1AZX0LIUHCPzYe8eg56w6W09TaxkWTnItET8+OZ1thFW0D2Kp0mxRU1QG43ZbTeNfJ4ag2R2G4Cg4SvrxwLFuPVPLxwQqvvObeYzU8u6mIbz25lT++s98rycbRpry63dmasRpI75RUN/LXD/JZPjOTS6eN4LnNRTS1Onp83tt7SokKC+acMUkAzMyOp6axlcMV9b4OuYMn3UdbRORFEbnRagq+17E6qnUfDWtXz80mJSaMP797wCuvt6nAuXPtBRNS+eXre7nvxZ39rlmsz6+gvLaJ4CBhx9GB7dce6n77Vh6tDuU7l0zgX+aNpLK+hTd3uV/4WVV5Z08Z541LISI0GIDpWQnAwM5X8CQpJAHHce7JbDUFH8svryciNIj02Ah/h2J8KCI0mFvPG8N7+8rY6YUP3E0FJ0iODuPhW87i9gvG8si6w9yxehONLT3/ddqdV7YfJTI0mGUzMth5tDqg51hsPHyC2qZWj849WFbLk58e4fpzchiVHM1541LIjI/gqQ3uu5D2ldRSVNnA4kkn9xebkB5DeEgQ2waw2OzJKqmn1xOspuBDh47XMTo52tbLCQA3zB9FTHgIf3n3YL9fa0tBJbNzEgkKEr5/+WR+eMVkXt1+jJtXraeqoffzIlodbby24xiLJ6cxb1QilfUtHK3yT7G5xdHGtsJKVn2Qz9cf38Ti/13L23sGboXaHUVVfP7PH3HLqvUeJdn739hHeEgQdyweDzi7C6+em817eWUcdTPnoH0o6kUTTyaFkOAgpmbGDa6Wgohki8hzIlIqIiUi8oyIZA9EcIHoUHmdFZkDRHxkKNefk8PL245ypB99xifqmjlYXsecUQkdx25bOJbfrpzFpoITXPvAOo718gPd2XXUzLLpGUzJjAdgZ9HA/LXa1qas3VvKL1/fw7/8ZR3T73ud5X/4kP96eRebD5/gRH0zf3h7/4DEAvD4+gJCg4WNBSe4c/Vmt91yW49U8sr2Ym5bOJbU2JMbVF4zbySq8MzG7gcCvLOnlCkZcYyIP7WXYEZ2AjuKqgdsCLMn3UcP4dwAJxPIAl5yHTO90OJo4+EP86lz0wRtdbRRUFHPmFRLCoHii+eNIThI+L/3+95a2HzEWU+Yc9raSlfOyuKhW86m8EQ9K/70IftLPd9X+uXtxUSFBbNoYhqTM2IRgZ29LDY3t7b1etRMQ7ODrz66kVse+pQH3j1IQ4uDlWfl8IfrZ/PRPYv56N6L+cbi8WwqqGTHACSp2qZWXthcxJWzsviPZVN4Y1cJ//HCji670lSV/35tD0nRYXx54ZhTHhuZFMW5ucn8Y2Nhl+9JVX0LGwtOnNJ11G5GdjwNLQ72l9Z67wdzw5OkkKqqD6lqq+v2MDB8N0f2kXUHjnPfS7v47Vt53Z5TVNlAa5vaFpwBZER8BJ+bncVTG45wvLapT6+x6XAlwUHCjOz4Mx47f3wKT35lAc0O5V8f9WzCXHvX0cWT04kMCyYqLISxKdG9Tgq3Pryei+5fy2ZXEbwnpTWNrHxwHWt2l/DDKyaz/b5LePEb53Pf8qksm5HZsezL1XOziQwN5u/rfL+wwotbjlLX7OC6s3O49bwxfPXCXB77pKDLlsr7eeV8dOA437hoHLERoWc8fu28kRRU1PNJ/pkjzt7NK8PRph1DUTubke1sAW4doC4kT5JCuYjcICLBrtsNOAvPphfyXFn+4Q8PddtVYKujBqbbL8ilqbWNRz461Kfnbz5ygkkjYokK63rR42lZ8dy3fAp5pbW8uLWox9dbd/A4FXXNXDE945TX2NWLgnhNYwvrDhznSEU9Vz+wjt++mUero63b8/eV1PC5P37EvpJaHrxxHrctHNvtzxMfGcpVszN5YWsRVfW+XUfq8fWHmTQiljk5zg/mu5dOZMXsLO5fs4+nPj1ZOG5rc7YSshMj+cL8nC5fa+m0EcRGhPCPLgrO7+wpJSk6jFkjE854bGxKNDHhIQO23IUnSeGLwLXAMaAYuNp1zPTCgbJaosOCCQqC/35tT5fnHLI5CgFpXFoMSyan88i6w267F7viaFO2FFSe0XV0usunZTA5I45fr8mjxc2HM8Ar24qJDgtm0cSTHQJTM+M4WtXIibpmj+LaePgEbQp/vH4On52Rwa/f3Me1f1nH4eN1Z5z74f5yPv/nj2h2tPHUVxZ0bHfqzo3zR9PY0ubxpLC+2FZYyY6iaq4/J6djRVwR4b+vnsHC8Snc+9z2joL3S9uOsvNoNd+5ZALhIcFdvl5EaDBXzsrk1R3FpyyK6HDVUC6ckNrlemdBQcK0rIErNnsy+qhAVZeraqqqpqnqVbYgXu/tL61lUkYcty8cy8vbijvGlXeWX15HdFgwqTHhXbyCGc6+uiiXqoYWVq8v6NXz9pXUUNfsYHbOmX9hdhYUJHz30gkUVNTzjw3dFztbHG28tvMYn5mS3jFWHmBqe7HZwy6k9fkVhAQJF05M5TcrZ/PblbPIK63l8t++zz82HOnok3/q0yPcvGo9mfGRPP/185jeRRdYV6ZkxjFvVCJ///iwz2b7Pv5JAZGhwVw1O+uU46HBQfz5hrlMyYjja49t4tNDFdz/xj4mjYjlyplZ3bya07XzRtLY0sZLW492HNty5AQn6lu67DpqNzM7gd3FNTS3uk/o3tBtUhCR/3Bz+3efRzbMHCyrZVxqDLdfmEtKTDg/fWX3GcWq/OP1jE6JdrtOvxme5uQkcvaYJP76QX6vfvE3F1R2PL8nF01MY05OAr97K6/boZUfHThOZX3LKV1H4GwpAB7PqVifX8H07PiOLqArZ2Xx2jcvYFpWPN99ehtff3wTP//nbr73zDYW5Cbzj39dQFYvl4q/ccEoDh+v5728sl49zxM1jS28uPUon52ZQVwX9YGY8BBW3XIWabERrHzwYwoq6rl76aQeh5JPz4pn0ohYnuqUmN/eU0pwkHDh+O5LtdOz42l2tLH3mOeDBfrKXUuhrosbwJeAu30c17BSWd9MeW0z49JiiAkP4TuXTGDj4RO8tuPUBdFsddTA9q8X5lJc1ciLnf6K7MmmghMkRYcxKrnnLkcR4a5LJ3KsupFHP+66sf/KtqPEhIdwwYRTP6ASosLISohkhwcthcYWB1sLKznbtVRDu6yESB7/8nzuuWwSa3aV8Jd3D3Ld2SNZdctZXX7w9uSyaRmkxIT7pOD8/Jaj1Dc7uP6cUd2ekxobziNfPJuEyFDOzU0+pbutOyLCtfNGsvVIZccH/Nt7ypibk0h8VPfvwUxXsXnbACyO121SUNX722/AgzhXS70VeAIY6/PIhpH2oWS5ac4P/GvmZjMhPYZfvLan46/C5tY2Ck/U28ijALZoYiqTRsTy4HsHPJ49vKngBLNHJnjcujw3N4XzxiXz57UHzpih29zaxus7S1hyWtdRuymZcR61FDYXVNLi0I71ezoLDhK+emEuL37jfH5/3Wx+9rnphAZ7Uto8U1hIENedPZK395b2a57H6VSVxz8pYEpGHDN76M4akxLN2u8uYtUtZ3n8b3DV7CxCg4WnNhyhuKqB3cXVLJ7cfdcRQHZiJIlRoWw74vtic0+rpCaJyE+AbTj3c56jqnerqvtFPMwpDpQ5k8K41FjAOUvx3ssnc/h4fcdfbEdO1NOmtuZRIBMRvnjeGPaV1LL5SM9/EVbWN3OwrI45o3q39/Ndl0zkeF0zD32Qf8rxDw+UU9XQwrIZGV0+b2pmHPnldT0Ww9fnVyDiXCa8O5Mz4vjszMx+d5Vef04OQSI89knvajHubC2sYndxNdd1KjC7ExsR2mUS7U5SdBhLpqTz3OYi3tjpLFR3NT+hMxFhenYC2wZgboa7msIvgU+BGmC6qt6nqp4NODan2F9aS1hIEFmJJ/tMF01IZeH4FH73dh5V9S2dRh5ZUghkl00fQURoEM9u6nkJ7PbE0VOR+XSzcxL5zOR0Hnz/4ClDOl/ZVkxsRAjnj0/p8nlTM+NRhT3H3HchrT90nEkj4oiP7H2XUG9lxEeyZHI6T35a0K91njp7/JPDRIUFc9WsTK+8XleumTeSirpmfvPmPrISIhmfFtPjc2ZkxbOvpIaGZu/8nN1x11L4Ds5ZzD8EjopItetWIyK2jm4vHCirY2xK9CnDzUSEey+bTFVDC394J69jjoK1FAJbbEQol04dwUtbi3tcannz4RMEycn+5t74ziUTqG1q5S/vOVdpbWp18PrOY1wyZUS3QypPFpu7//VvcbSx6XBll11HvnLjglGcqG/hlW3F/X6t6sYWXtpazPKZmV1OQPOWC8anMiIughP1LSyelOZRi2RGdjyONmVXsW8/ft3VFIJUNVJVY1U1rtMtVlXjfBrVMLO/tJZxXfwlMCUzjmvmZvPIR4d5P6+cuIgQEt0Um0xgWDEnm6qGFt7Z476XdlNBJRNHxBEd3vUkL3cmZ8Tx2RmZPPThIUprGvkgr5yaxtZuu44AMuIjSIwKZWdR9x9KO4qqaGhxnFFk9qVzc5PJTY3mb90Uz3vj+c1FNLQ4uP6criegeUv7InnQc9dRu/aZzb6er9C3Co/xWGOLgyMn6slN7bp5+O0lEwkOEt7dV8YYG45qgPNyk0mLDeeZTd3PPna0KVuOVHbMtO2Lby2ZQLOjjT+9c4BXthUTFxHCeeO67joCZ+t2amY8O4u779de71rC4azRA5cURIQb549i65HKfn1gtheYp2XFdXwA+9JtC8dwz2WTWNhNd93pRsRHkBYb7vOZzX5JCiKSICJPi8geEdktIgtcRe01IpLn+tq76tkglV9ehypdthTA+Q/95Qucg7msnmDAORDhqtlZvLOnlIpuZhDvL62ltqnVo/kJ3RmTEs3Vc7J5/JMCXt95jEunjiAsxP1HwtTMOPYdq+12VvT6/ArGpkafskLoQFgxN5uosGD+1o/hqZsKKtlzrIbrz+5+GKo3JUSF8dULcwnpxeirGdnxPl8DyV8thd8Cr6nqJGAmsBu4B3hLVccDb7nuD3kdw1G7aSkAfOWCsUxMj+Xc3OSBCssMcivmZNHapry4pevWQvuM+N4WmU9352eca/7XNTu4wk3XUbspmXE0O9rIKzlzxU5Hm7L+UMWA1hPaxUWE8rnZWby09ajHS3Gc7vFPCogOC2a5DwvM/TUjO4GD5XXUNPpuzafed0b2k4jEARcAtwCoajPQLCJXAotcpz0CrGUYTJI7UFaLCIx1sxx2dHgIr3/rggGMygx2k0bEMTUzjmc3F3HLeWPOeHzT4RMkRoX2e2BCVkIkt54/mpe2HHXbddSufbmLHUermJJ5amlx77EaahpbB7Se0NlNC0bz2CcFfPfpbczIjic6PITosGCiwkOICXeu9hoWEkSrQ2l1tNHS5vrqUJodbby87Sgr5mQT04cazUCZke0cAbajqJoFPvoj0h8//VigDHhIRGYCG4F/A9JVtRhAVYtFpMvqi4jcDtwOkJPj22KQN+wvrWVkYlSvxjEbA86C849f3kVeSQ3j02NPeWzzEedOa96oQd2zdBLfWTLRo0lkY1KiiQwNZlcXI5DW5zsXTx7IekJnE0fEcsWMDNbsKuHN3b3fmU0EbuhmhdPBor3Wsb2oclglhRBgDnCHqn4iIr+lF11FqvogzhnWzJs3b9BvGru/tJZc2zTH9MHymZn87NXdPLu5iLuXTuo4XlXfwv7SWq6c6Z1uDhEhLMSz5BIcJEzOiO1yZvP6QxVkJUSSnei/VX7/eP0cwDk0tr7JQV1zK3VNrdQ1O6hvaqXJ0UZoUBAhwUJosBAcFERIkBAaHER8ZOgZu54NNknRYWQnRrLVh8VmfySFQqBQVT9x3X8aZ1IoEZEMVyshAxjys6YdbUp+eZ3HowuM6Sw1NpwLJ6Ty/OYi7rpkYsc8l46d1no5k9lbpmXF84xrB7H2BeBUlfX5JwbN//XQ4CDio4Lcric0VM3IjvfpsNQBLzSr6jHgiIhMdB26GNiFc8vPm13HbgZeGOjYvK3oRANNrW1ui8zGuLNiThbFVY18fPDkvlabCyoRgZldbMgyEKZmxlHX7OBwp/WG8svrKK9t8ls9IZDMyE7gSEVDnwvqPfHX6KM7gMdEZBswC/gZ8AtgiYjkAUtc94e0jjWPPJjCbkxXPjM5ndiIkFM2fN9UcIKJ6bF+K4ie3FvhZBdG+/wESwq+NyPL+f77ah0kvyQFVd2iqvNUdYZr054TqnpcVS9W1fGur2duZDrEeDIc1Rh3IkKDWTYjk3/uOEZdUyttrklrs/sxP6G/xqfHEBIkpyx3sT6/gpSYMMbaXBufm+ZauXW7j7qQbEazD+0vrSU5OozE6DB/h2KGsM/PyaKhxcFrO46xv6yWmsbWfs1k7q/wkGDGp8eekhQ+ya/g7DFJNiN/AMRFhDI2JdpnxebBOyB3GDhQVkuudR2Zfpo7KpGcpCie3VxIa5tzJrE/WwrgrCus3VuKqlJU2UBRZQNfXnjmfArjG3/8whzS43wzUsqSgkt1Ywtr95bxxs5jvLevjCmZcdx1yUTm9XHMtaqyv6yWy6b1PEvUGHdEhBVzsvjtW3kEiRAfGer3bpqpmXE8vbGQ0pomPj3kWu/I6gkDZnKG79YkDeikUFrdyJrdJby+s4R1B8ppcSgpMWFcPDmd9/PKufqBdSyamMpdl0xkWpZnG4q3q6hrprK+xYrMxitWzM7mN2/m8X5eOYsmpva4F7CvdcxsLqpifX4FsREhTBphiycPBwGZFHYereKHz+/o2PR8VHIUt543hkumpDM7J5HgIKGh2cEj6w7xwLsHWPb7D7hs2gi+vWTCGTNLu9NeZLakYLwhJzmKs0cnsf5QRb8WwfOWyRnO34OdR6tZn1/BWaOTTtkvxAxdAZkUUmPCaXUo31kygUumjmBCeswZBbLIsGC+emEu15+Tw1/fz+f/vX+Q13ce46rZWXznkolkJUR28+pO+8vaRx7ZaAzjHSvmZLH+UAVz/TRprbPYiFBGJ0fx3r4yDpTVcc28kf4OyXhJQCaFtLgIXrrjfI/OjYsI5VtLJnDzuaN54N0DPPLRIbYXVvHGty5wO9LiQGkdkaHBZMa7Tx7GeOqaeSNJiwsfNKvpTs2M55Xtzt3ObH7C8GFDUj2UFB3G9y+fzI+vmkZeaS0bDrvfrnp/WS1jU6P93vdrho/gIGHxpPRBM+yzfZXUyNBgpmX2ruZmBi9LCr20bEYGseEhrF5f4Pa8A91swWnMcNG+Z/OcUQk9bs5jhg77l+ylqLAQls/K5NXtxVQ1dL3RRX1zK0WVDYyzmcxmGJuWFU9wkLBg7ODozjLeYUmhD1aelUNjSxsvdLMr1sGyOgCbuGaGtZSYcJ7/2nnctnCsv0MxXmRJoQ+mZ8czNTOO1euPoHrmlg62EJ4JFNOz420DqWHGkkIfrTw7h93F1WzvYqXC/aW1BIlz/oMxxgwllhT66MpZmUSEBrF6/ZEzHjtQVsuo5GjCQ+wvKGPM0GJJoY/iIkK5YnomL24poq6p9ZTHnFtwWteRMWbosaTQD9edPZK6ZgevbCvuONbqaCO/vI7cNJvJbIwZeiwp9MPcUYmMS4th9acn5ywcOdFAi0NtOKoxZkiypNAPIsLKs0ayuaCSvcdqgE67rdnII2PMEGRJoZ9WzMkmLDiIJ1ythQNltgWnMWbosqTQT0nRYVwyNZ3nNhfR2OJgf2ktqbHhxEeG+js0Y4zpNUsKXnDd2TlU1rfw+s5j7C+ttXqCMWbIsqTgBQvGJjMyKZLV6ws4UGYL4Rljhi5LCl4QFCSsPCuHjw9WUNPYahvrGGOGLEsKXnLN3OyO7QjHpXm2Zacxxgw2lhS8JC0ugsWT0gBs4poxZsgKyO04feW7l05kckYcI+Ii/B2KMcb0iSUFL5qQHsu3l1jXkTFm6PJb95GIBIvIZhF52XU/SUTWiEie62uiv2IzxphA5c+awr8Buzvdvwd4S1XHA2+57htjjBlAfkkKIpINXAH8v06HrwQecX3/CHDVQMdljDGBzl8thd8A3wPaOh1LV9ViANfXtK6eKCK3i8gGEdlQVlbm+0iNMSaADHhSEJFlQKmqbuzL81X1QVWdp6rzUlNTvRydMcYENn+MPjoPWC4ilwMRQJyIPAqUiEiGqhaLSAZQ6ofYjDEmoA14S0FV71XVbFUdDawE3lbVG4AXgZtdp90MvDDQsRljTKAbTDOafwEsEZE8YInrvjHGmAEkqurvGPpMRMqAw/14iRSg3EvhDEf2/rhn70/P7D1yz1/vzyhV7bIoO6STQn+JyAZVnefvOAYre3/cs/enZ/YeuTcY35/B1H1kjDHGzywpGGOM6RDoSeFBfwcwyNn74569Pz2z98i9Qff+BHRNwRhjzKkCvaVgjDGmE0sKxhhjOgRkUhCRpSKyV0T2i4gt0Q2IyCoRKRWRHZ2O2R4XLiIyUkTeEZHdIrJTRP7NddzeI0BEIkRkvYhsdb0//+k6bu9PJ0NhH5mASwoiEgz8EbgMmAJcJyJT/BvVoPAwsPS0Y7bHxUmtwHdUdTIwH/i66/+NvUdOTcBiVZ0JzAKWish87P053aDfRybgkgJwNrBfVQ+qajPwBM69HAKaqr4HVJx22Pa4cFHVYlXd5Pq+Bucvdhb2HgGgTrWuu6Gum2LvT4ehso9MICaFLOBIp/uFrmPmTB7tcRFoRGQ0MBv4BHuPOri6RrbgXOF4jara+3OqPu8jM5ACMSlIF8dsXK7xiIjEAM8A31TVan/HM5ioqkNVZwHZwNkiMs3fMQ0W/d1HZiAFYlIoBEZ2up8NHPVTLINdiWtvC2yPCxCRUJwJ4TFVfdZ12N6j06hqJbAWZ43K3h+n9n1kDuHssl7ceR8ZGDzvTyAmhU+B8SIyRkTCcO7p8KKfYxqsbI8LFxER4K/AblX9VaeH7D0CRCRVRBJc30cCnwH2YO8PMLT2kQnIGc2uXd9+AwQDq1T1p34Oye9EZDWwCOdSviXAj4DngaeAHKAAuEZVTy9GBwQROR94H9jOyT7h7+OsKwT8eyQiM3AWSoNx/rH5lKr+l4gkY+/PKURkEXCXqi4bjO9PQCYFY4wxXQvE7iNjjDHdsKRgjDGmgyUFY4wxHSwpGGOM6WBJwRhjTAdLCsYAIpIsIltct2MiUuT6vlZE/uSja35TRG5y8/iy9tVGjRkoNiTVmNOIyH1Arar+rw+vEQJsAuaoams354jrnPNUtd5XsRjTmbUUjHFDRBZ1Wvv+PhF5RETeEJFDIrJCRP5HRLaLyGuuZTAQkbki8q6IbBSR19uXMTjNYmBTe0IQkTtFZJeIbBORJ8C58ijO5SKWDcgPawyWFIzprVycyx9fCTwKvKOq04EG4ApXYvg9cLWqzgVWAV3NmD8P6Lw42j3AbFWdAXy10/ENwEKv/xTGdCPE3wEYM8T8U1VbRGQ7ziUdXnMd3w6MBiYC04A1zt4fgoHiLl4ng1M3W9kGPCYiz+NcXqRdKZDpzR/AGHcsKRjTO00AqtomIi16sijXhvP3SYCdqrqgh9dpACI63b8CuABYDvy7iEx1dS1FuM41ZkBY95Ex3rUXSBWRBeBcbltEpnZx3m5gnOucIGCkqr6DcxOWBCDGdd4EYEcXzzfGJywpGONFri1erwb+W0S2AluAc7s49Z84Wwbg7GJ61NUltRn4tWtPAoCLgFd8G7UxJ9mQVGP8RESeA76nqnndPJ4OPK6qFw9sZCaQWVIwxk9EZCLOPXrf6+bxs4AWVd0ysJGZQGZJwRhjTAerKRhjjOlgScEYY0wHSwrGGGM6WFIwxhjTwZKCMcaYDv8fQXu+QP858/cAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.title('AIF')        \n",
    "#y = normalize_zero_one(aif)\n",
    "#y_p = normalize_zero_one(vof)\n",
    "#plt.plot(mu_voi)\n",
    "#plt.plot(y_p)\n",
    "plt.plot(label)\n",
    "plt.legend([ 'AIF','VOF'])\n",
    "plt.xlabel('Time (s)')\n",
    "plt.ylabel('Normalized Density (HU)')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "d995139f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.utils.data import DataLoader\n",
    "\n",
    "train_dataloader = DataLoader(training_data, batch_size=1, shuffle=True)\n",
    "#test_dataloader = DataLoader(test_data, batch_size=64, shuffle=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "28b1748e",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_features, train_labels = next(iter(train_dataloader))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "74904fc9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 256, 256, 8, 43])"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_features.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "4479324a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 43])"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "519fb609",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}