475_Final_Project.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/alexathomases/covid19-classification/blob/main/475_Final_Project.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vlGrQfDuycJx"
      },
      "source": [
        "#Downloading Dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RF99dJtwl4kN",
        "outputId": "ac18fb2a-3ea1-4b60-cfef-4928f1774cdf"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "# mount Google Drive (must do every time to access saved files)\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vPzVQ6QytCGg"
      },
      "outputs": [],
      "source": [
        "#Configure Kaggle API\n",
        "import os\n",
        "os.environ['KAGGLE_USERNAME'] = \"alexholzbach\" # username from the kaggle json file\n",
        "os.environ['KAGGLE_KEY'] = \"2d7d33825c5a175807651729ce5e80d3\" # key from the kaggle json file"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PSV9cGIRt8ED"
      },
      "outputs": [],
      "source": [
        "!cd drive/Shareddrives/'475 Final Project'/datasets"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Kw3yRkzrtlq9",
        "outputId": "8c0a0d15-ef74-4442-ef02-3a9143d6f691"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading large-covid19-ct-slice-dataset.zip to /content\n",
            "100% 2.05G/2.06G [00:18<00:00, 148MB/s]\n",
            "100% 2.06G/2.06G [00:18<00:00, 122MB/s]\n"
          ]
        }
      ],
      "source": [
        "# download dataset into the notebook\n",
        "!kaggle datasets download -d maedemaftouni/large-covid19-ct-slice-dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1UATSb-2wuH7"
      },
      "outputs": [],
      "source": [
        "# store dataset in shared drive\n",
        "!mv '/content/large-covid19-ct-slice-dataset.zip' '/content/drive/Shareddrives/475 Final Project/datasets'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "qqnzNgddzB42"
      },
      "outputs": [],
      "source": [
        "!unzip '/content/drive/Shareddrives/475 Final Project/datasets/large-covid19-ct-slice-dataset.zip'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "gzSWKRVHz5Zy"
      },
      "outputs": [],
      "source": [
        "!mv '/content/curated_data' '/content/drive/Shareddrives/475 Final Project/datasets'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "XWi9dqqR5TAN"
      },
      "outputs": [],
      "source": [
        "!mv '/content/meta_data_normal.csv' '/content/drive/Shareddrives/475 Final Project/datasets'"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Azx1eFa6ykGR"
      },
      "source": [
        "#Start Code"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Be3qKJKskKao",
        "outputId": "6f70f7a8-d9eb-4255-e9fe-6c97de37380c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting progressbar\n",
            "  Downloading progressbar-2.5.tar.gz (10 kB)\n",
            "Building wheels for collected packages: progressbar\n",
            "  Building wheel for progressbar (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for progressbar: filename=progressbar-2.5-py3-none-any.whl size=12082 sha256=26b51a27b56eef74bf0643fd11d34a83cd61fda6f7d6e68486dfce93f5d55d07\n",
            "  Stored in directory: /root/.cache/pip/wheels/f0/fd/1f/3e35ed57e94cd8ced38dd46771f1f0f94f65fec548659ed855\n",
            "Successfully built progressbar\n",
            "Installing collected packages: progressbar\n",
            "Successfully installed progressbar-2.5\n"
          ]
        }
      ],
      "source": [
        "import pandas as pd\n",
        "from pathlib import Path\n",
        "import os\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from matplotlib import pyplot as plt, image as img\n",
        "import numpy as np\n",
        "from sklearn.model_selection import train_test_split\n",
        "from tensorflow.keras.layers import BatchNormalization\n",
        "from tensorflow.keras.layers import (\n",
        "    Conv2D,\n",
        "    Conv2DTranspose,\n",
        "    MaxPool2D,\n",
        "    ReLU,\n",
        "    Flatten,\n",
        "    Dense,\n",
        "    UpSampling2D,\n",
        "    Dropout,\n",
        "    Input, \n",
        "    Concatenate,\n",
        "    GlobalAveragePooling2D, \n",
        "    GlobalMaxPooling2D,\n",
        ")\n",
        "from skimage.transform import resize\n",
        "from tensorflow import Tensor\n",
        "from tensorflow.keras.models import Model, Sequential, load_model\n",
        "from tensorflow.keras import layers\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.callbacks import Callback, ModelCheckpoint, ReduceLROnPlateau, EarlyStopping\n",
        "from keras.layers import Input\n",
        "from tensorflow import Tensor\n",
        "from scipy.stats import mannwhitneyu\n",
        "from tabulate import tabulate\n",
        "\n",
        "!pip3 install progressbar\n",
        "import progressbar\n",
        "\n",
        "# random seed generator\n",
        "np.random.seed(3)\n",
        "tf.random.set_seed(7)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 236
        },
        "id": "r7xnmkVTMePE",
        "outputId": "94b67181-1b1a-401f-86e7-1018f8532fc9"
      },
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-5-6128681ff2a0>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mmetadata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mread_csv\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mencoding\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'cp720'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mmetadata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdrop_duplicates\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"Patient ID\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minplace\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0mfilenames\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmetadata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"File name\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mfilepaths\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mpath\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m\"/\"\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mimg\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mimg\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mfilenames\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfilepaths\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'path' is not defined"
          ]
        }
      ],
      "source": [
        "metadata = pd.read_csv(path, encoding='cp720')\n",
        "metadata.drop_duplicates([\"Patient ID\"], inplace=True)\n",
        "filenames = np.array(metadata[\"File name\"])\n",
        "filepaths = [path + \"/\" + img for img in filenames]\n",
        "print(filepaths)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fzTEKgd1MGr8"
      },
      "outputs": [],
      "source": [
        "\n",
        "def get_disctinct_filepaths(metadatapath, imgdirpath):\n",
        "  metadata = pd.read_csv(metadatapath, encoding=\"cp720\")\n",
        "  metadata.drop_duplicates([\"Patient ID\"], inplace=True)\n",
        "  filenames = np.array(metadata[\"File name\"])\n",
        "  filepaths = [imgdirpath + \"/\" + img for img in filenames]\n",
        "\n",
        "  num_patients = len(metadata[\"Patient ID\"])\n",
        "  print(f\"Found {num_patients} unique patients at {imgdirpath}\")\n",
        "  return filepaths"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "eZo4127k-33J"
      },
      "outputs": [],
      "source": [
        "def load_raw_data():\n",
        "  #filepaths for metadata\n",
        "  metacovpath = \"/content/drive/Shareddrives/475 Final Project/datasets/meta_data_covid.csv\"\n",
        "  metanonpath = \"/content/drive/Shareddrives/475 Final Project/datasets/meta_data_normal.csv\"\n",
        "  metacappath = \"/content/drive/Shareddrives/475 Final Project/datasets/meta_data_cap.csv\"\n",
        "\n",
        "  #filepaths for images\n",
        "  imgcovpath = \"/content/drive/Shareddrives/475 Final Project/datasets/curated_data/curated_data/2COVID\"\n",
        "  imgnonpath = \"/content/drive/Shareddrives/475 Final Project/datasets/curated_data/curated_data/1NonCOVID\"\n",
        "  imgcappath = \"/content/drive/Shareddrives/475 Final Project/datasets/curated_data/curated_data/3CAP\"\n",
        "\n",
        "  #distinct patients in each\n",
        "  covid_images = get_disctinct_filepaths(metacovpath, imgcovpath)\n",
        "  non_covid_images = get_disctinct_filepaths(metanonpath, imgnonpath)\n",
        "  cap_images = get_disctinct_filepaths(metacappath, imgcappath)\n",
        "  \n",
        "  #Showing imagess\n",
        "  fig = plt.figure(figsize=(10, 10))\n",
        "  fig.add_subplot(1, 2, 1)\n",
        "  image = img.imread(covid_images[1])\n",
        "  plt.imshow(image)\n",
        "  plt.title(\"CT Scan of Lungs with COVID-19\")\n",
        "\n",
        "  fig.add_subplot(1, 2, 2)\n",
        "  image = img.imread(non_covid_images[1])\n",
        "  plt.imshow(image)\n",
        "  plt.title(\"CT Scan of Normal Lungs\")\n",
        "  return covid_images, non_covid_images, cap_images"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XzAiwrQG_ULG",
        "outputId": "afddfaf6-ca95-420f-d3fe-b2aff3bec768"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 464 unique patients at /content/drive/Shareddrives/475 Final Project/datasets/curated_data/curated_data/2COVID\n",
            "Found 604 unique patients at /content/drive/Shareddrives/475 Final Project/datasets/curated_data/curated_data/1NonCOVID\n",
            "Found 54 unique patients at /content/drive/Shareddrives/475 Final Project/datasets/curated_data/curated_data/3CAP\n"
          ]
        }
      ],
      "source": [
        "covid_images, non_covid_images, cap_images = load_raw_data()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QnhgHsGv_nMF"
      },
      "outputs": [],
      "source": [
        "IMG_SIZE = 256\n",
        "\n",
        "# Three empty numpy arrays to store coverted images\n",
        "positive_npy = np.empty((len(covid_images), IMG_SIZE, IMG_SIZE, 1), dtype=np.float32)\n",
        "negative_npy = np.empty((len(non_covid_images), IMG_SIZE, IMG_SIZE, 1), dtype=np.float32)\n",
        "cap_npy = np.empty((len(cap_images), IMG_SIZE, IMG_SIZE, 1), dtype=np.float32)\n",
        "\n",
        "# start a bar of show percentage of loading data \n",
        "# covid_bar = progressbar.ProgressBar(maxval=len(covid_images), widgets=[progressbar.Bar('=', '[', ']'), ' ', progressbar.Percentage()])\n",
        "# non_covid_bar = progressbar.ProgressBar(maxval=len(non_covid_images), widgets=[progressbar.Bar('=', '[', ']'), ' ', progressbar.Percentage()])\n",
        "# cap_bar = progressbar.ProgressBar(maxval=len(cap_images), widgets=[progressbar.Bar('=', '[', ']'), ' ', progressbar.Percentage()])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        },
        "id": "JWzY1iIXO6Gr",
        "outputId": "ec744385-0713-4d82-a11a-746662cb3c3c"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f325eea5490>"
            ]
          },
          "execution_count": 8,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "test_img = img.imread(covid_images[4])\n",
        "test_resize = resize(test_img, (IMG_SIZE, IMG_SIZE), anti_aliasing=True)\n",
        "plt.imshow(test_resize)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XKcCSJfR_sgl",
        "outputId": "23cb4a54-dc70-4bfa-ae4f-3e6a4c023841"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "0.0\n",
            "0.04310344827586207\n",
            "0.08620689655172414\n",
            "0.12931034482758622\n",
            "0.1724137931034483\n",
            "0.21551724137931033\n",
            "0.25862068965517243\n",
            "0.3017241379310345\n",
            "0.3448275862068966\n",
            "0.3879310344827586\n",
            "0.43103448275862066\n",
            "0.47413793103448276\n",
            "0.5172413793103449\n",
            "0.5603448275862069\n",
            "0.603448275862069\n",
            "0.646551724137931\n",
            "0.6896551724137931\n",
            "0.7327586206896551\n",
            "0.7758620689655172\n",
            "0.8189655172413793\n",
            "0.8620689655172413\n",
            "0.9051724137931034\n",
            "0.9482758620689655\n",
            "0.9913793103448276\n",
            "COVID images converting done\n"
          ]
        }
      ],
      "source": [
        "# covid_bar.start()\n",
        "# Converting COVID dataset to .npy format\n",
        "k = len(covid_images)\n",
        "for i, _file in enumerate(covid_images):\n",
        "    try:\n",
        "        image_npy = img.imread(_file)\n",
        "        positive = resize(image_npy, (IMG_SIZE, IMG_SIZE, 1), anti_aliasing=True)\n",
        "        positive_npy[i] = positive\n",
        "    except:\n",
        "        pass\n",
        "    if i%20 == 0:\n",
        "      print(i/k)\n",
        "    # covid_bar.update(i+1)\n",
        "# covid_bar.finish()\n",
        "print(\"COVID images converting done\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HA3YC8e8_422",
        "outputId": "78370d80-320d-4d3c-82b8-98f725dc758b"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "0.0\n",
            "0.033112582781456956\n",
            "0.06622516556291391\n",
            "0.09933774834437085\n",
            "0.13245033112582782\n",
            "0.16556291390728478\n",
            "0.1986754966887417\n",
            "0.23178807947019867\n",
            "0.26490066225165565\n",
            "0.2980132450331126\n",
            "0.33112582781456956\n",
            "0.36423841059602646\n",
            "0.3973509933774834\n",
            "0.4304635761589404\n",
            "0.46357615894039733\n",
            "0.4966887417218543\n",
            "0.5298013245033113\n",
            "0.5629139072847682\n",
            "0.5960264900662252\n",
            "0.6291390728476821\n",
            "0.6622516556291391\n",
            "0.695364238410596\n",
            "0.7284768211920529\n",
            "0.7615894039735099\n",
            "0.7947019867549668\n",
            "0.8278145695364238\n",
            "0.8609271523178808\n",
            "0.8940397350993378\n",
            "0.9271523178807947\n",
            "0.9602649006622517\n",
            "0.9933774834437086\n",
            "non covid images converting done\n"
          ]
        }
      ],
      "source": [
        "# non_covid_bar.start()\n",
        "k = len(non_covid_images)\n",
        "# Converting non-COVID dataset to .npy format\n",
        "for i, _file in enumerate(non_covid_images):\n",
        "    try:\n",
        "        image_npy = img.imread(_file)\n",
        "        negative = resize(image_npy, (IMG_SIZE, IMG_SIZE, 1), anti_aliasing=True)\n",
        "        negative_npy[i] = negative\n",
        "    except:\n",
        "        print(\"cant open image \" + i + \"th\") \n",
        "    # non_covid_bar.update(i+1)\n",
        "    if i%20 == 0:\n",
        "      print(i/k)\n",
        "# non_covid_bar.finish()\n",
        "print(\"non covid images converting done\")     "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lD6ltxayA-yC",
        "outputId": "b0332dc4-35e2-4c3a-8c6e-a207c4f04134"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "0.0\n",
            "0.37037037037037035\n",
            "0.7407407407407407\n",
            "cap images converting done\n"
          ]
        }
      ],
      "source": [
        "# cap_bar.start()\n",
        "# Converting non-COVID dataset to .npy format\n",
        "k = len(cap_images)\n",
        "for i, _file in enumerate(cap_images):\n",
        "    try:\n",
        "        image_npy = img.imread(_file)\n",
        "        cap = resize(image_npy, (IMG_SIZE, IMG_SIZE, 1), anti_aliasing=True)\n",
        "        cap_npy[i] = cap\n",
        "    except:\n",
        "        print(\"cant open image \" + i + \"th\") \n",
        "    if i%20 == 0:\n",
        "      print(i/k)\n",
        "    # cap_bar.update(i+1)\n",
        "# cap_bar.finish()\n",
        "print(\"cap images converting done\") "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Kw4CSoYmBGoq",
        "outputId": "70d0bfc1-eb66-4921-e751-7265d49f9ffb"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "(464, 256, 256, 1)\n",
            "(604, 256, 256, 1)\n",
            "(54, 256, 256, 1)\n"
          ]
        }
      ],
      "source": [
        "print(positive_npy.shape)\n",
        "print(negative_npy.shape)\n",
        "print(cap_npy.shape)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "3ogVtSNyBMaL"
      },
      "outputs": [],
      "source": [
        "# To load .npy datasets\n",
        "def load_data():\n",
        "    capnum = 11\n",
        "    posnum = 93\n",
        "    negnum = 121\n",
        "    cap = cap_npy\n",
        "    cap_labels = [2 for i in cap]\n",
        "    cap_test = cap[:capnum,:,:,:]\n",
        "    cap_train = cap[capnum:,:,:,:]\n",
        "    positive = positive_npy\n",
        "    positive_labels = [0 for i in positive]\n",
        "    positive_test = positive[:posnum,:,:,:]\n",
        "    positive_train = positive[posnum:,:,:,:]\n",
        "    negative = negative_npy\n",
        "    negative_labels = [1 for i in negative]\n",
        "    negative_test = negative[:negnum,:,:,:]\n",
        "    negative_train = negative[negnum:,:,:,:]\n",
        "\n",
        "    # Joining both datasets and labels\n",
        "    X_train = np.concatenate([positive_train, negative_train, cap_train])\n",
        "    X_test = np.concatenate([positive_test, negative_test, cap_test])\n",
        "    y_train = np.concatenate([positive_labels[posnum:], negative_labels[negnum:], cap_labels[capnum:]])\n",
        "    y_test = np.concatenate([positive_labels[:posnum], negative_labels[:negnum], cap_labels[:capnum]])\n",
        "    return X_train, X_test, y_train, y_test"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "TUZUCRcoBRjo"
      },
      "outputs": [],
      "source": [
        "# save x and y\n",
        "X_train, X_test, y_train, y_test = load_data()\n",
        "np.savez('X_train', X_train)\n",
        "np.savez('X_test', X_test)\n",
        "np.savez('y_train', y_train)\n",
        "np.savez('y_test', y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OLVxp3Kkss8M"
      },
      "outputs": [],
      "source": [
        "!mv '/content/y_test.npz' '/content/drive/Shareddrives/475 Final Project'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vqUdr2pknnqe"
      },
      "outputs": [],
      "source": [
        "# load in \n",
        "dataX_train = np.load('/content/drive/Shareddrives/475 Final Project/X_train.npz')\n",
        "dataX_test = np.load('/content/drive/Shareddrives/475 Final Project/X_test.npz')\n",
        "datay_train = np.load('/content/drive/Shareddrives/475 Final Project/y_train.npz')\n",
        "datay_test = np.load('/content/drive/Shareddrives/475 Final Project/y_test.npz')\n",
        "X_train = dataX_train['arr_0']\n",
        "X_test = dataX_test['arr_0']\n",
        "y_train = datay_train['arr_0']\n",
        "y_test = datay_test['arr_0']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PWy3hZB5BUV4"
      },
      "outputs": [],
      "source": [
        "# print(X.shape)\n",
        "# print(y.shape)\n",
        "\n",
        "# X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n",
        "y_train_hot = keras.utils.to_categorical(y_train, 3)\n",
        "y_test_hot =  keras.utils.to_categorical(y_test, 3)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "p6XYbugRBmCp",
        "outputId": "02b98cd8-7198-4b90-d7fc-103288e2fe1a"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg19/vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "80142336/80134624 [==============================] - 1s 0us/step\n",
            "80150528/80134624 [==============================] - 1s 0us/step\n",
            "0 input_1\n",
            "1 block1_conv1\n",
            "2 block1_conv2\n",
            "3 block1_pool\n",
            "4 block2_conv1\n",
            "5 block2_conv2\n",
            "6 block2_pool\n",
            "7 block3_conv1\n",
            "8 block3_conv2\n",
            "9 block3_conv3\n",
            "10 block3_conv4\n",
            "11 block3_pool\n",
            "12 block4_conv1\n",
            "13 block4_conv2\n",
            "14 block4_conv3\n",
            "15 block4_conv4\n",
            "16 block4_pool\n",
            "17 block5_conv1\n",
            "18 block5_conv2\n",
            "19 block5_conv3\n",
            "20 block5_conv4\n",
            "21 block5_pool\n",
            "<keras.engine.input_layer.InputLayer object at 0x7f3262bdc3d0> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f325eed4b10> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f325b5af990> False\n",
            "<keras.layers.pooling.MaxPooling2D object at 0x7f325acef110> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f3250785a50> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f3250788a50> False\n",
            "<keras.layers.pooling.MaxPooling2D object at 0x7f325078cc10> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f3250794590> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f325079d2d0> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f3250791a90> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f32507a6810> False\n",
            "<keras.layers.pooling.MaxPooling2D object at 0x7f32507b0910> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f32507aab10> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f32507b43d0> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f325073ba90> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f3250740a90> False\n",
            "<keras.layers.pooling.MaxPooling2D object at 0x7f3250745210> False\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f3250749e50> True\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f3250753310> True\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f325074e610> True\n",
            "<keras.layers.convolutional.Conv2D object at 0x7f325075a410> True\n",
            "<keras.layers.pooling.MaxPooling2D object at 0x7f325076a110> True\n",
            "Model: \"vgg19\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_1 (InputLayer)        [(None, None, None, 3)]   0         \n",
            "                                                                 \n",
            " block1_conv1 (Conv2D)       (None, None, None, 64)    1792      \n",
            "                                                                 \n",
            " block1_conv2 (Conv2D)       (None, None, None, 64)    36928     \n",
            "                                                                 \n",
            " block1_pool (MaxPooling2D)  (None, None, None, 64)    0         \n",
            "                                                                 \n",
            " block2_conv1 (Conv2D)       (None, None, None, 128)   73856     \n",
            "                                                                 \n",
            " block2_conv2 (Conv2D)       (None, None, None, 128)   147584    \n",
            "                                                                 \n",
            " block2_pool (MaxPooling2D)  (None, None, None, 128)   0         \n",
            "                                                                 \n",
            " block3_conv1 (Conv2D)       (None, None, None, 256)   295168    \n",
            "                                                                 \n",
            " block3_conv2 (Conv2D)       (None, None, None, 256)   590080    \n",
            "                                                                 \n",
            " block3_conv3 (Conv2D)       (None, None, None, 256)   590080    \n",
            "                                                                 \n",
            " block3_conv4 (Conv2D)       (None, None, None, 256)   590080    \n",
            "                                                                 \n",
            " block3_pool (MaxPooling2D)  (None, None, None, 256)   0         \n",
            "                                                                 \n",
            " block4_conv1 (Conv2D)       (None, None, None, 512)   1180160   \n",
            "                                                                 \n",
            " block4_conv2 (Conv2D)       (None, None, None, 512)   2359808   \n",
            "                                                                 \n",
            " block4_conv3 (Conv2D)       (None, None, None, 512)   2359808   \n",
            "                                                                 \n",
            " block4_conv4 (Conv2D)       (None, None, None, 512)   2359808   \n",
            "                                                                 \n",
            " block4_pool (MaxPooling2D)  (None, None, None, 512)   0         \n",
            "                                                                 \n",
            " block5_conv1 (Conv2D)       (None, None, None, 512)   2359808   \n",
            "                                                                 \n",
            " block5_conv2 (Conv2D)       (None, None, None, 512)   2359808   \n",
            "                                                                 \n",
            " block5_conv3 (Conv2D)       (None, None, None, 512)   2359808   \n",
            "                                                                 \n",
            " block5_conv4 (Conv2D)       (None, None, None, 512)   2359808   \n",
            "                                                                 \n",
            " block5_pool (MaxPooling2D)  (None, None, None, 512)   0         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 20,024,384\n",
            "Trainable params: 0\n",
            "Non-trainable params: 20,024,384\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "VGG = tf.keras.applications.VGG19(weights='imagenet', include_top=False)\n",
        "VGG.trainable =False\n",
        "for i, layer in enumerate(VGG.layers):\n",
        "  print(i, layer.name)\n",
        "\n",
        "for layer in VGG.layers[:-5]:\n",
        "   layer.trainable = False\n",
        "for layer in VGG.layers[-5:]:\n",
        "   layer.trainable = True\n",
        "for layer in VGG.layers:\n",
        "    print(layer, layer.trainable) \n",
        "\n",
        "VGG.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "6Lg4rDyuBsH1"
      },
      "outputs": [],
      "source": [
        "def cnn(): \n",
        "    from tensorflow.keras.layers import Input, GlobalAveragePooling2D, GlobalMaxPooling2D, Concatenate\n",
        "    from tensorflow.keras.models import Model\n",
        "    VGG = tf.keras.applications.VGG19(weights='imagenet', include_top=False)\n",
        "    for layer in VGG.layers[:-5]:\n",
        "      layer.trainable = False\n",
        "    for layer in VGG.layers[-5:]:\n",
        "      layer.trainable = True\n",
        "    \n",
        "    input = Input(shape=(IMG_SIZE, IMG_SIZE, 1))\n",
        "    x = Conv2D(3, (3, 3), padding='same')(input)\n",
        "    \n",
        "    x = VGG(x , training=False)\n",
        "    \n",
        "    x = GlobalAveragePooling2D()(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = Dropout(0.5)(x)\n",
        "    x = BatchNormalization()(x)\n",
        "    x = Dropout(0.5)(x)\n",
        "\n",
        "    # multi output\n",
        "    output = Dense(3, activation=\"softmax\", name='root')(x)\n",
        " \n",
        "\n",
        "    # model\n",
        "    model = Model(input,output)\n",
        "    \n",
        "    optimizer = keras.optimizers.Adam(lr=0.002, beta_1=0.9, beta_2=0.999, epsilon=0.1, decay=0.0)\n",
        "    model.compile(loss='categorical_crossentropy', optimizer=optimizer, metrics=['accuracy'])\n",
        "    \n",
        "    return model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iOeVVwYyByTk",
        "outputId": "ad42b21d-e8a7-417e-8631-6fdda25beaf5"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_3 (InputLayer)        [(None, 256, 256, 1)]     0         \n",
            "                                                                 \n",
            " conv2d (Conv2D)             (None, 256, 256, 3)       30        \n",
            "                                                                 \n",
            " vgg19 (Functional)          (None, None, None, 512)   20024384  \n",
            "                                                                 \n",
            " global_average_pooling2d (G  (None, 512)              0         \n",
            " lobalAveragePooling2D)                                          \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 512)              2048      \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 512)               0         \n",
            "                                                                 \n",
            " batch_normalization_1 (Batc  (None, 512)              2048      \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 512)               0         \n",
            "                                                                 \n",
            " root (Dense)                (None, 3)                 1539      \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 20,030,049\n",
            "Trainable params: 9,442,849\n",
            "Non-trainable params: 10,587,200\n",
            "_________________________________________________________________\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/optimizer_v2/adam.py:105: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
            "  super(Adam, self).__init__(name, **kwargs)\n"
          ]
        }
      ],
      "source": [
        "model = cnn()\n",
        "model.summary()\n",
        "\n",
        "annealer = ReduceLROnPlateau(monitor='val_accuracy', factor=0.5, patience=5, verbose=1, min_lr=1e-3)\n",
        "checkpoint = ModelCheckpoint('model.h5', verbose=1, save_best_only=True)\n",
        "\n",
        "datagen = ImageDataGenerator(rotation_range=180, # Degree range for random rotations\n",
        "                        width_shift_range=0.2, # Range for random horizontal shifts\n",
        "                        height_shift_range=0.2, # Range for random vertical shifts\n",
        "                        zoom_range=0.2, # Range for random zoom\n",
        "                        horizontal_flip=True, # Randomly flip inputs horizontally\n",
        "                        vertical_flip=True) # Randomly flip inputs vertically\n",
        "BATCH_SIZE = 64\n",
        "datagen.fit(X_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "background_save": true,
          "base_uri": "https://localhost:8080/"
        },
        "id": "KUQkVDZvB2yO",
        "outputId": "df2748b9-a3e7-45de-d066-c2014db301c6"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "15/15 [==============================] - ETA: 0s - loss: 1.2074 - accuracy: 0.5619  \n",
            "Epoch 1: val_loss improved from inf to 0.80035, saving model to model.h5\n",
            "15/15 [==============================] - 2019s 134s/step - loss: 1.2074 - accuracy: 0.5619 - val_loss: 0.8004 - val_accuracy: 0.6578 - lr: 0.0020\n",
            "Epoch 2/20\n",
            "12/15 [=======================>......] - ETA: 6:30 - loss: 0.8163 - accuracy: 0.7357"
          ]
        }
      ],
      "source": [
        "history = model.fit(X_train, y_train_hot, batch_size=BATCH_SIZE,\n",
        "               epochs=20,\n",
        "               verbose=1,\n",
        "               callbacks=[annealer, checkpoint],\n",
        "               validation_data=(X_test, y_test_hot))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "LcsZYjltB7QC",
        "outputId": "9bc1a9d9-4fd6-439c-ed12-d17070db8581"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# accuracy plot \n",
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "a28YDprYCHMv",
        "outputId": "1cbc6304-54c3-4c61-b439-1e520ab20f62"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# loss plot\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "6-6Dw5AexWhQ"
      },
      "outputs": [],
      "source": [
        "y_pred = model.predict(X_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0NJR_iT3x4ey"
      },
      "outputs": [],
      "source": [
        "from sklearn import metrics\n",
        "matrix = metrics.confusion_matrix(y_test, y_pred.argmax(axis=1))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 501
        },
        "id": "y6IqndA3ybAA",
        "outputId": "802946d8-22a9-4243-de26-da58680756ae"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 540x540 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "fig, ax = plt.subplots(figsize=(7.5, 7.5))\n",
        "ax.matshow(matrix, cmap=plt.cm.Blues, alpha=0.3)\n",
        "for i in range(matrix.shape[0]):\n",
        "    for j in range(matrix.shape[1]):\n",
        "        ax.text(x=j, y=i,s=matrix[i, j], va='center', ha='center', size='xx-large')\n",
        " \n",
        "plt.xlabel('Predictions', fontsize=18)\n",
        "plt.ylabel('Actuals', fontsize=18)\n",
        "plt.title('Confusion Matrix', fontsize=18)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bnWOgvfqSCr5"
      },
      "outputs": [],
      "source": [
        "!mv '/content/model.h5' '/content/drive/Shareddrives/475 Final Project'"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "10hQvLQWSEdx"
      },
      "source": [
        "# Load Trained Model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ygZuqucESECV"
      },
      "outputs": [],
      "source": [
        "model = keras.models.load_model('/content/drive/Shareddrives/475 Final Project/model.h5')\n",
        "model.summary()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZfBqBFkCVtgH"
      },
      "source": [
        "# GradCam Visualization"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ffndMc5eVsJf"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "\n",
        "from IPython.display import Image\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.cm as cm"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bg0ZUsm6H_sP"
      },
      "outputs": [],
      "source": [
        "[a.name for a in model.layers[2].layers]\n",
        "\n",
        "#model.layers[2].get_layer('block4_conv1').output"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "EOohLITpVnRl"
      },
      "outputs": [],
      "source": [
        "# https://keras.io/examples/vision/grad_cam/\n",
        "\n",
        "# The Grad-CAM algorithm\n",
        "def get_img_array(img_path, size):\n",
        "    # `img` is a PIL image of size 299x299\n",
        "    img = keras.preprocessing.image.load_img(img_path, target_size=size)\n",
        "    # `array` is a float32 Numpy array of shape (299, 299, 3)\n",
        "    array = keras.preprocessing.image.img_to_array(img)\n",
        "    # We add a dimension to transform our array into a \"batch\"\n",
        "    # of size (1, 64, 64, 3)\n",
        "    array = np.expand_dims(array, axis=0)\n",
        "    return array\n",
        "\n",
        "def make_gradcam_heatmap(img_array, model):\n",
        "    final_conv_layer = model.get_layer('vgg19').get_layer('block5_conv3')\n",
        "\n",
        "    # Model connecting input image to final conv layer\n",
        "    input = keras.Input(shape=model.inputs[0].shape[1:])\n",
        "    output = input\n",
        "    output = model.layers[0](input)\n",
        "    output = model.layers[1](output)\n",
        "    for layer in model.get_layer('vgg19').layers[:-1]: # slice to -1 to exclude final pooling layer\n",
        "         output = layer(output)\n",
        "    input_to_conv_model = keras.Model(input, output)\n",
        "\n",
        "    # Model connecting final conv layer to predictions\n",
        "    input = keras.Input(shape = final_conv_layer.output.shape[1:]) # 1: cuts out batch dimension\n",
        "    output = input\n",
        "    # pass through final pooling layer\n",
        "    output = model.get_layer('vgg19').layers[-1](input)\n",
        "    # iterate through layers after vgg\n",
        "    for layer in model.layers[3:]: # 3: to slice after vgg19 layer\n",
        "        output = layer(output)\n",
        "    conv_to_predict_model = keras.Model(input, output)\n",
        "\n",
        "    # Compute the gradient of the top predicted class for our input image\n",
        "    # with respect to the activations of the last conv layer\n",
        "    with tf.GradientTape() as tape:\n",
        "        # Compute activations of the last conv layer and make the tape watch it\n",
        "        conv_output = input_to_conv_model(img_array)\n",
        "        tape.watch(conv_output)\n",
        "        # Compute class predictions\n",
        "        preds = conv_to_predict_model(conv_output)\n",
        "        top_pred_index = tf.argmax(preds[0])\n",
        "        top_class_channel = preds[:, top_pred_index]\n",
        "\n",
        "    # This is the gradient of the top predicted class with regard to\n",
        "    # the output feature map of the last conv layer\n",
        "    grads = tape.gradient(top_class_channel, conv_output)\n",
        "\n",
        "    # This is a vector where each entry is the mean intensity of the gradient\n",
        "    # over a specific feature map channel\n",
        "    pooled_grads = tf.reduce_mean(grads, axis=(0, 1, 2))\n",
        "\n",
        "    # We multiply each channel in the feature map array\n",
        "    # by \"how important this channel is\" with regard to the top predicted class\n",
        "    conv_output = conv_output.numpy()[0]\n",
        "    pooled_grads = pooled_grads.numpy()\n",
        "    for i in range(pooled_grads.shape[-1]):\n",
        "        conv_output[:, :, i] *= pooled_grads[i]\n",
        "\n",
        "    # The channel-wise mean of the resulting feature map\n",
        "    # is our heatmap of class activation\n",
        "    heatmap = np.mean(conv_output, axis=-1)\n",
        "\n",
        "    # For visualization purpose, we will also normalize the heatmap between 0 & 1\n",
        "    heatmap = np.maximum(heatmap, 0) / np.max(heatmap)\n",
        "    return heatmap"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "rqY9EKP2uDI5"
      },
      "outputs": [],
      "source": [
        "# Test Gradcam\n",
        "k = 65\n",
        "image_path = non_covid_images[k]\n",
        "image_size = (IMG_SIZE, IMG_SIZE)\n",
        "img_array = get_img_array(image_path, size=image_size)\n",
        "img_array = tf.image.rgb_to_grayscale(img_array)\n",
        "#print(y_test[k], y_pred.argmax(axis=1)[k])\n",
        "# Remove last layer's softmax\n",
        "model.layers[-1].activation = None\n",
        "\n",
        "# Generate class activation heatmap\n",
        "heatmap = make_gradcam_heatmap(img_array, model)\n",
        "\n",
        "#Display heatmap\n",
        "plt.matshow(heatmap)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 529
        },
        "id": "Q-2kD1N4y7D0",
        "outputId": "41d94f41-5c76-4cd6-8e50-b1d6188826b9"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Image object>"
            ],
            "image/jpeg": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "def save_and_display_gradcam(img_path, heatmap, cam_path=\"cam.jpg\", alpha=0.4):\n",
        "    # Load the original image\n",
        "    img = keras.preprocessing.image.load_img(img_path)\n",
        "    img = keras.preprocessing.image.img_to_array(img)\n",
        "\n",
        "    # Rescale heatmap to a range 0-255\n",
        "    heatmap = np.uint8(255 * heatmap)\n",
        "\n",
        "    # Use jet colormap to colorize heatmap\n",
        "    jet = cm.get_cmap(\"jet\")\n",
        "\n",
        "    # Use RGB values of the colormap\n",
        "    jet_colors = jet(np.arange(256))[:, :3]\n",
        "    jet_heatmap = jet_colors[heatmap]\n",
        "\n",
        "    # Create an image with RGB colorized heatmap\n",
        "    jet_heatmap = keras.preprocessing.image.array_to_img(jet_heatmap)\n",
        "    jet_heatmap = jet_heatmap.resize((img.shape[1], img.shape[0]))\n",
        "    jet_heatmap = keras.preprocessing.image.img_to_array(jet_heatmap)\n",
        "\n",
        "    # Superimpose the heatmap on original image\n",
        "    superimposed_img = jet_heatmap * alpha + img\n",
        "    superimposed_img = keras.preprocessing.image.array_to_img(superimposed_img)\n",
        "\n",
        "    # Save the superimposed image\n",
        "    superimposed_img.save(cam_path)\n",
        "\n",
        "    # Display Grad CAM\n",
        "    display(Image(cam_path))\n",
        "\n",
        "save_and_display_gradcam(covid_images[3], heatmap)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Z_ha73E5j8yx"
      },
      "source": [
        "# Perform PCA"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 219
        },
        "id": "2QN2Cxftj_Ga",
        "outputId": "4ad9b6f3-7162-4bcb-db35-2ca257f4b4b3"
      },
      "outputs": [
        {
          "ename": "NameError",
          "evalue": "ignored",
          "output_type": "error",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-1-534d66e1fcf4>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdecomposition\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mPCA\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0mpca\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mPCA\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mn_components\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mX_reduced\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpca\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit_transform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_train\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'X_train' is not defined"
          ]
        }
      ],
      "source": [
        "from sklearn.decomposition import PCA\n",
        "print(X_train.shape)\n",
        "pca = PCA()\n",
        "X_reduced = pca.fit_transform(X_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2N0pcYuQ6CGH"
      },
      "outputs": [],
      "source": [
        "BATCH_SIZE = 64\n",
        "datagen.fit(X_reduced)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0XlS1izc6g-d"
      },
      "outputs": [],
      "source": [
        "history = model.fit(X_reduced, y_train_hot, batch_size=BATCH_SIZE,\n",
        "               epochs=50,\n",
        "               verbose=1,\n",
        "               callbacks=[annealer, checkpoint],\n",
        "               validation_data=(X_test, y_test_hot))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "H38qbdaw6vIX"
      },
      "outputs": [],
      "source": [
        "# accuracy plot \n",
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5ci2gotm62uH"
      },
      "outputs": [],
      "source": [
        "# loss plot\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'test'], loc='upper left')\n",
        "plt.show()"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "collapsed_sections": [],
      "name": "475_Final_Project.ipynb",
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}