Pneumonia_vs_Normal.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Pneumonia vs Normal.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyPL5DY7SwoOlE7Ch14r1oB0",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/daeIy/Normal-vs-Pneumonia/blob/master/Pneumonia_vs_Normal.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WF8tqJUWslpu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "! pip install -q kaggle"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "faJ-SHBOtSHu",
        "colab_type": "code",
        "colab": {
          "resources": {
            "http://localhost:8080/nbextensions/google.colab/files.js": {
              "data": "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",
              "ok": true,
              "headers": [
                [
                  "content-type",
                  "application/javascript"
                ]
              ],
              "status": 200,
              "status_text": ""
            }
          },
          "base_uri": "https://localhost:8080/",
          "height": 89
        },
        "outputId": "bd07861a-c3bf-4a1f-899e-a437f8998976"
      },
      "source": [
        "# Upload Kaggle API key\n",
        "\n",
        "from google.colab import files\n",
        "files.upload()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "     <input type=\"file\" id=\"files-d7e422f6-0902-4407-b2ca-45a74d2c17fd\" name=\"files[]\" multiple disabled\n",
              "        style=\"border:none\" />\n",
              "     <output id=\"result-d7e422f6-0902-4407-b2ca-45a74d2c17fd\">\n",
              "      Upload widget is only available when the cell has been executed in the\n",
              "      current browser session. Please rerun this cell to enable.\n",
              "      </output>\n",
              "      <script src=\"/nbextensions/google.colab/files.js\"></script> "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "Saving kaggle.json to kaggle.json\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'kaggle.json': b'{\"username\":\"gabrieldaely\",\"key\":\"26bf1049d09710bdea490aa8549a8716\"}'}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3ynU0Ht0tXOp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "! mkdir ~/.kaggle\n",
        "! cp kaggle.json ~/.kaggle/\n",
        "! chmod 600 ~/.kaggle/kaggle.json"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bICRzwwKtcUj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "d84b8b3c-b86c-4825-a2ce-8eac28e611c7"
      },
      "source": [
        "! kaggle datasets download -d 'paultimothymooney/chest-xray-pneumonia'"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading chest-xray-pneumonia.zip to /content\n",
            "100% 2.29G/2.29G [00:29<00:00, 32.6MB/s]\n",
            "100% 2.29G/2.29G [00:29<00:00, 82.8MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OYStKtATteHM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "45afe5ad-7f0b-4ee3-f164-bcd5c380143a"
      },
      "source": [
        "! ls"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "chest-xray-pneumonia.zip  kaggle.json  sample_data\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f8lwAHu4uxor",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "! unzip -q chest-xray-pneumonia.zip\n",
        "! rm -r chest-xray-pneumonia.zip"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A5WK26Xcu0ns",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from tensorflow.keras.applications import DenseNet121\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import MaxPooling2D\n",
        "from tensorflow.keras.layers import Dropout\n",
        "from tensorflow.keras.layers import Flatten\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.layers import Input\n",
        "from tensorflow.keras.models import Model\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import os"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SrkOpUXvxXoJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        },
        "outputId": "e062bfc9-51e6-4ab8-f035-9c0fcc8347a2"
      },
      "source": [
        "print(len(os.listdir('./chest_xray/train/NORMAL')))\n",
        "print(len(os.listdir('./chest_xray/train/PNEUMONIA')))\n",
        "print(len(os.listdir('./chest_xray/test/NORMAL')))\n",
        "print(len(os.listdir('./chest_xray/test/PNEUMONIA')))\n",
        "print(len(os.listdir('./chest_xray/val/NORMAL')))\n",
        "print(len(os.listdir('./chest_xray/val/PNEUMONIA')))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "1341\n",
            "3875\n",
            "234\n",
            "390\n",
            "8\n",
            "8\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fdm7G9mnyCb7",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "3cd9a046-78ed-4417-ab17-cb7259c9b19c"
      },
      "source": [
        "TRAIN_DIR = './chest_xray/train'\n",
        "VAL_DIR = './chest_xray/test'\n",
        "\n",
        "# Create train data generator\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale = 1./255,\n",
        "    rotation_range=40,\n",
        "    width_shift_range=0.2,\n",
        "    height_shift_range=0.2,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True,\n",
        "    fill_mode='nearest')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "\tTRAIN_DIR,\n",
        "\ttarget_size=(224,224),\n",
        "\tclass_mode='binary',\n",
        "    batch_size=32\n",
        ")\n",
        "\n",
        "# Create validation data generator\n",
        "val_datagen = ImageDataGenerator(rescale = 1./255)\n",
        "\n",
        "val_generator = val_datagen.flow_from_directory(\n",
        "\tVAL_DIR,\n",
        "\ttarget_size=(224,224),\n",
        "\tclass_mode='binary',\n",
        "    batch_size=32\n",
        ")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 5216 images belonging to 2 classes.\n",
            "Found 624 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IqGdd22vybM2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "densenet = DenseNet121(weights=\"imagenet\", include_top=False, input_tensor=Input(shape=(224, 224, 3)))\n",
        "\n",
        "for layer in densenet.layers:\n",
        "\tlayer.trainable = False\n",
        "\n",
        "mod1 = densenet.output\n",
        "mod1 = MaxPooling2D()(mod1)\n",
        "mod1 = Flatten()(mod1)\n",
        "mod1 = Dense(64, activation=\"relu\")(mod1)\n",
        "mod1 = Dropout(0.5)(mod1)\n",
        "mod1 = Dense(1, activation=\"sigmoid\")(mod1)\n",
        "\n",
        "model = Model(inputs=densenet.input, outputs=mod1)\n",
        "model.compile(loss = 'binary_crossentropy', optimizer='adam', metrics=['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JI4BdZ6E1r7f",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "23c2c54e-efbd-47cd-ab40-d08f93b884cd"
      },
      "source": [
        "history = model.fit(train_generator, epochs = 30, \n",
        "                    validation_data = val_generator,\n",
        "                    verbose = 1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/30\n",
            "163/163 [==============================] - 131s 806ms/step - loss: 0.4866 - accuracy: 0.8376 - val_loss: 0.5402 - val_accuracy: 0.7260\n",
            "Epoch 2/30\n",
            "163/163 [==============================] - 128s 785ms/step - loss: 0.2562 - accuracy: 0.8972 - val_loss: 0.3670 - val_accuracy: 0.8205\n",
            "Epoch 3/30\n",
            "163/163 [==============================] - 128s 788ms/step - loss: 0.2340 - accuracy: 0.9045 - val_loss: 0.4864 - val_accuracy: 0.7821\n",
            "Epoch 4/30\n",
            "163/163 [==============================] - 130s 796ms/step - loss: 0.2501 - accuracy: 0.9114 - val_loss: 0.3098 - val_accuracy: 0.8878\n",
            "Epoch 5/30\n",
            "163/163 [==============================] - 129s 793ms/step - loss: 0.2207 - accuracy: 0.9151 - val_loss: 0.2820 - val_accuracy: 0.9038\n",
            "Epoch 6/30\n",
            "163/163 [==============================] - 129s 794ms/step - loss: 0.2372 - accuracy: 0.8959 - val_loss: 0.3881 - val_accuracy: 0.8686\n",
            "Epoch 7/30\n",
            "163/163 [==============================] - 128s 785ms/step - loss: 0.2396 - accuracy: 0.9149 - val_loss: 0.4240 - val_accuracy: 0.8894\n",
            "Epoch 8/30\n",
            "163/163 [==============================] - 128s 788ms/step - loss: 0.2307 - accuracy: 0.9222 - val_loss: 0.4549 - val_accuracy: 0.8894\n",
            "Epoch 9/30\n",
            "163/163 [==============================] - 129s 791ms/step - loss: 0.2325 - accuracy: 0.9162 - val_loss: 0.3331 - val_accuracy: 0.9054\n",
            "Epoch 10/30\n",
            "163/163 [==============================] - 129s 793ms/step - loss: 0.2070 - accuracy: 0.9254 - val_loss: 0.3439 - val_accuracy: 0.8942\n",
            "Epoch 11/30\n",
            "163/163 [==============================] - 129s 793ms/step - loss: 0.2114 - accuracy: 0.9225 - val_loss: 0.2687 - val_accuracy: 0.9103\n",
            "Epoch 12/30\n",
            "163/163 [==============================] - 128s 787ms/step - loss: 0.1826 - accuracy: 0.9289 - val_loss: 0.3487 - val_accuracy: 0.9103\n",
            "Epoch 13/30\n",
            "163/163 [==============================] - 129s 790ms/step - loss: 0.1963 - accuracy: 0.9227 - val_loss: 0.3291 - val_accuracy: 0.8910\n",
            "Epoch 14/30\n",
            "163/163 [==============================] - 130s 795ms/step - loss: 0.1830 - accuracy: 0.9293 - val_loss: 0.2637 - val_accuracy: 0.9135\n",
            "Epoch 15/30\n",
            "163/163 [==============================] - 130s 798ms/step - loss: 0.1894 - accuracy: 0.9281 - val_loss: 0.3062 - val_accuracy: 0.8990\n",
            "Epoch 16/30\n",
            "163/163 [==============================] - 129s 792ms/step - loss: 0.1791 - accuracy: 0.9310 - val_loss: 0.3668 - val_accuracy: 0.9038\n",
            "Epoch 17/30\n",
            "163/163 [==============================] - 129s 791ms/step - loss: 0.1864 - accuracy: 0.9264 - val_loss: 0.2785 - val_accuracy: 0.9071\n",
            "Epoch 18/30\n",
            "163/163 [==============================] - 130s 796ms/step - loss: 0.1713 - accuracy: 0.9369 - val_loss: 0.3314 - val_accuracy: 0.9038\n",
            "Epoch 19/30\n",
            "163/163 [==============================] - 129s 793ms/step - loss: 0.1819 - accuracy: 0.9327 - val_loss: 0.2989 - val_accuracy: 0.9022\n",
            "Epoch 20/30\n",
            "163/163 [==============================] - 130s 795ms/step - loss: 0.1642 - accuracy: 0.9354 - val_loss: 0.3347 - val_accuracy: 0.9119\n",
            "Epoch 21/30\n",
            "163/163 [==============================] - 129s 789ms/step - loss: 0.1742 - accuracy: 0.9342 - val_loss: 0.2684 - val_accuracy: 0.9151\n",
            "Epoch 22/30\n",
            "163/163 [==============================] - 129s 789ms/step - loss: 0.1753 - accuracy: 0.9314 - val_loss: 0.3178 - val_accuracy: 0.9135\n",
            "Epoch 23/30\n",
            "163/163 [==============================] - 128s 785ms/step - loss: 0.1707 - accuracy: 0.9352 - val_loss: 0.3290 - val_accuracy: 0.9135\n",
            "Epoch 24/30\n",
            "163/163 [==============================] - 125s 768ms/step - loss: 0.1711 - accuracy: 0.9342 - val_loss: 0.2594 - val_accuracy: 0.9087\n",
            "Epoch 25/30\n",
            "163/163 [==============================] - 128s 788ms/step - loss: 0.1841 - accuracy: 0.9273 - val_loss: 0.3419 - val_accuracy: 0.9103\n",
            "Epoch 26/30\n",
            "163/163 [==============================] - 129s 789ms/step - loss: 0.1703 - accuracy: 0.9352 - val_loss: 0.2652 - val_accuracy: 0.9183\n",
            "Epoch 27/30\n",
            "163/163 [==============================] - 130s 795ms/step - loss: 0.1664 - accuracy: 0.9371 - val_loss: 0.4065 - val_accuracy: 0.9038\n",
            "Epoch 28/30\n",
            "163/163 [==============================] - 129s 791ms/step - loss: 0.1714 - accuracy: 0.9319 - val_loss: 0.3209 - val_accuracy: 0.9263\n",
            "Epoch 29/30\n",
            "163/163 [==============================] - 129s 790ms/step - loss: 0.1665 - accuracy: 0.9340 - val_loss: 0.2989 - val_accuracy: 0.9135\n",
            "Epoch 30/30\n",
            "163/163 [==============================] - 129s 794ms/step - loss: 0.1566 - accuracy: 0.9415 - val_loss: 0.3447 - val_accuracy: 0.9119\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ej4S5piTN1Va",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 336
        },
        "outputId": "96a551a5-f79d-4e63-cbf9-f4b3f8bdac2d"
      },
      "source": [
        "acc = history.history['accuracy']\n",
        "val_acc = history.history['val_accuracy']\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs = range(len(acc))\n",
        "\n",
        "plt.figure(figsize=(10,5))\n",
        "plt.subplot(121)\n",
        "plt.plot(epochs, acc, 'r', label='Training accuracy')\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation accuracy')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.subplot(122)\n",
        "plt.plot(epochs, loss, 'r', label='Training Loss')\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation Loss')\n",
        "plt.ylim(0,2)\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P0DXGb7mRhFd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.save('pneumonia_vs_normal9491.h5')"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}